OP-JCRE190043 1..24

People Rely Less on Consumer Reviews for

Experiential than Material Purchases

HENGCHEN DAI

CINDY CHAN

CASSIE MOGILNER

An increasingly prevalent form of social inﬂuence occurs online where consumers

read reviews written by other consumers. Do people rely on consumer reviews differ-

ently when making experiential purchases (events to live through) versus when mak-

ing material purchases (objects to keep)? Though people often use consumer

reviews both when making experiential and material purchases, an analysis of more

than six million reviews on Amazon.com and four laboratory experiments reveal that

people are less likely to rely on consumer reviews for experiential purchases than for

material purchases. This effect is driven by beliefs that reviews are less reﬂective of

the purchase’s objective quality for experiences than for material goods. These ﬁnd-

ings not only indicate how different types of purchases are inﬂuenced by word of

mouth, but also illuminate the psychological processes underlying shoppers’ reliance

on consumer reviews. Furthermore, as one of the ﬁrst investigations into how people

choose among various experiential and material purchase options, these ﬁndings

suggest that people are less receptive to being told what to do than what to have.

Keywords: experiential purchases, material purchases, consumer reviews,

objective quality

n preparing for her first ski trip, this article’s first author

needs to book a hotel room and make restaurant reserva-

tions, and also get a camera and a pair of skis. While all of

these purchases will be instrumental to her enjoyment of

the trip, the former are more experiential (events to live

through), whereas the latter are more material (possessions

to keep). How might this difference in purchase type affect

her decision process? Will she rely on consumer reviews

equally when choosing a hotel and restaurants and when

choosing a camera and skis?

It is well known that people’s attitudes toward products,

services, and retailers are often shaped by others (Cialdini

and Goldstein 2004; Goldstein, Cialdini, and Griskevicius

2008), and an increasingly prevalent form of social influ-

ence occurs online where people read reviews written by

other consumers (Chen and Xie 2008; Keen 2008;

Mayzlin, Dover, and Chevalier 2012). Recent surveys re-

port that more than 90% of people read consumer reviews

before making a purchase (Marchant 2015), and two-thirds

of people trust opinions of anonymous online consumers

(Nielsen 2015). It is thus not surprising that consumer

reviews can have a considerable influence on product sales

Hengchen Dai is an assistant professor of management and organizations,

the UCLA Anderson School of Management, University of California, 110

Westwood Plaza, Suite A405, Los Angeles, CA, 90095 (hengchen.dai@

anderson.ucla.edu; 310-206-2716); Cindy Chan is an assistant professor of mar-

keting, UTSC Department of Management and Rotman School of Management,

University of Toronto, 1095 Military Trail, Toronto, ON M1C 1A4, Canada

([email protected]; 416-978-1462); and Cassie Mogilner is an associate

professor of marketing, the UCLA Anderson School of Management, University

of California, 110 Westwood Plaza, Suite B515, Los Angeles, CA 90095

(cassie.holmes@anderson.ucla.edu; 310-794-7714). Correspondence: Hengchen

Dai. The authors thank Jeffrey Cai, Raghu Iyengar, Katherine Milkman, Katie

Shonk, and participants at the OPIM-DP lab at the Wharton School for their

helpful feedback on this article. The authors thank the Wharton Behavioral Lab,

the Anderson Behavioral Lab, Sheng Xu, Katie Wirtz, Sonia Fung, Kahini

Shah, Emily Short, and Mayoori Baskarasingham for their help with data collec-

tion. The authors thank Wharton Dean’s Research Fund, the Wharton Risk

Management and Decision Processes Center, Wharton’s Baker Retailing Center,

the UCLA Anderson School of Management, the Anderson Behavioral Lab,

Olin Business School, and the Department of Management at the University of

Toronto Scarborough for funding support. In addition to the fun from working

together on this project, the authors really appreciated the guidance gained from

consumer reviews to find the best quality ice cream makers available, as well as

the excuse to try out all their nearby ice cream shops to evaluate their goodness

for themselves! Supplementary materials are included in the web appendix ac-

companying the online version of this article.

Editors: Darren W. Dahl and Margaret C. Campbell

Associate Editor: Cait Lamberton

Advance Access publication September 10, 2019

The Author(s) 2019. Published by Oxford University Press on behalf of Journal of Consumer Research, Inc.

DOI: 10.1093/jcr/ucz042

Downloaded from https://academic.oup.com/jcr/advance-article-abstract/doi/10.1093/jcr/ucz042/5567101 by UCLA Digital Collections Services user on 20 February 2020

(Chevalier and Mayzlin 2006; Godes and Mayzlin 2004;

Moe and Trusov 2011) and even stock prices (Tirunillai

and Tellis 2012). Given the increasing importance of con-

sumer reviews, it has become critical to understand the ex-

tent to which people rely on this source of information

across their various purchase decisions.

The current research compares experiential and material

purchases and tests how and why people rely on consumer

reviews differently when making these purchases. By

“reliance on consumer reviews,” we mean the extent to

which people find the consumer reviews they read useful

and are influenced by them. Although there are countless

review websites with endless numbers of consumer

reviews that shoppers can search for and read, not all of

these reviews are equally helpful to consumers in deciding

what to buy. Indeed, many shopping and review web-

sites—such as Amazon, Best Buy, Expedia, and

TripAdvisor—allow shoppers to vote whether the con-

sumer reviews on their site are helpful or not (Forman,

Ghose, and Wiesenfeld 2008; Yin, Bond, and Zhang 2014,

2017). This research examines whether the extent to which

people consider consumer reviews helpful differs depend-

ing on the type of purchase being made.

Though people often use consumer reviews both when

making experiential purchases and when making material

purchases, we find across an archival field study and four

laboratory experiments that people rely less on consumer

reviews for experiential purchases than for material pur-

chases. This is because people believe that assessments of

experiences (compared to material goods) are based less on

the purchase’s objective quality, which makes other con-

sumers’ reviews less helpful for their purchase decision.

These findings make important contributions to three

streams of literature. First, prior work comparing experien-

tial and material purchases has predominantly focused on

the post-purchase effects of experiences relative to material

possessions (see Gilovich, Kumar, and Jampol 2015a for a

review). Though newer work has begun focusing on pre-

purchase effects by examining when and why people might

choose one purchase type over the other (Goodman and

Lim 2018; Goodman, Malkoc, and Stephenson 2016;

Kumar and Gilovich 2015, 2016; Kumar, Killingsworth,

and Gilovich 2014; Pchelin and Howell 2014; Tully,

Hershfield, and Meyvis 2015), our work advances this lit-

erature by examining how the decision processes differ

when people are deciding what to buy within each pur-

chase type. Second, these findings inform the field’s under-

standing of how different types of purchases are influenced

by word of mouth (Berger 2014). By identifying beliefs

about objective quality as a key driver, this research further

illuminates the psychological processes underlying the per-

ceived usefulness of consumer reviews (Chen and Lurie

2013; de Langhe, Fernbach, and Lichtenstein 2016; Moore

2015; Yin et al. 2014, 2017). Third, building on new re-

search revealing that individuals vary in their beliefs about

the extent to which purchases are assessed based on objec-

tive quality (Spiller and Belogolova 2017), our findings

highlight that these beliefs also systematically vary across

different purchase types. Furthermore, we document an im-

portant implication of such beliefs by demonstrating their

impact on people’s reliance on consumer reviews.

EXPERIENTIAL VERSUS MATERIAL

PURCHASES

Van Boven and Gilovich (2003) define experiential pur-

chases as “those made with the primary intention of acquir-

ing a life experience: an event or series of events that one

lives through” and material purchases as “those made with

the primary intention of acquiring a material good: a tangi-

ble object that is kept in one’s possession” (1194).

Although the two categories cannot always be precisely

separated, consumers share the intuition underlying this

classification and can readily place a purchase on the

experiential-material spectrum (Gilovich, Kumar, and

Jampol 2015b; Van Boven and Gilovich 2003).

To date, the research comparing experiential and mate-

rial purchases has largely focused on understanding the

consequences of these purchases, such as the happiness

(Van Boven and Gilovich 2003), satisfaction (Carter and

Gilovich 2010), gratitude (Walker, Kumar, and Gilovich

2016), and regret (Rosenzweig and Gilovich 2012) they

elicit, as well as their effect on interpersonal relationships

(Chan and Mogilner 2017). Research exploring the reasons

for these outcomes shows that compared to material pos-

sessions, experiential purchases are more closely tied to

one’s self-identity (Carter and Gilovich 2012), harder to

compare against forgone alternatives (Carter and Gilovich

2010), less interchangeable across options (Rosenzweig

and Gilovich 2012), subject to slower rates of hedonic ad-

aptation (Nicolao, Irwin, and Goodman 2009), more likely

to elicit intense emotions (Chan and Mogilner 2017), and

more often shared with others (Caprariello and Reis 2013;

Kumar and Gilovich 2015). For a review, see Gilovich

et al. (2015a).

More recently, research comparing experiential and ma-

terial purchases has begun examining differences that oc-

cur before a purchase is made. Researchers have found that

people derive greater utility from anticipating and talking

about future experiential purchases (Kumar et al. 2014;

Kumar and Gilovich 2015) and are therefore willing to

wait longer before consuming experiences than material

goods (Kumar and Gilovich 2016). Also, when deciding

between purchasing an experience versus a material good,

people prefer the material good when they feel financially

constrained (Tully et al. 2015), and they mistakenly fore-

cast that material goods are a better use of money (Pchelin

and Howell 2014). Further, people predict that material

goods will make better gifts (Goodman and Lim 2018), yet

2 JOURNAL OF CONSUMER RESEARCH

Downloaded from https://academic.oup.com/jcr/advance-article-abstract/doi/10.1093/jcr/ucz042/5567101 by UCLA Digital Collections Services user on 20 February 2020

view experiences as better for celebrating special life

events (Goodman et al. 2016).

Though researchers have started to explore factors that

influence when and why people might decide to make one

type of purchase over the other, research has yet to exam-

ine differences that exist between experiential and material

purchases in terms of how people decide what option to

buy within a given purchase type. Our investigation con-

tributes by examining whether and why people are differ-

entially influenced by others when making an experiential

purchase versus a material purchase. We specifically ex-

amine the extent to which people rely on consumer reviews

when making these two types of purchases, as well as how

people’s beliefs about the basis of assessment play a key

role.

BELIEFS ABOUT ASSESSMENTS OF

OBJECTIVE QUALITY

Though price and personal taste can play a role, a key

factor that contributes to a consumer’s overall assessment

of a purchase option is its objective quality (Johansson,

Douglas, and Nonaka 1985; Zeithaml 1988). Quality-based

assessments reflect consumers’ judgments about a prod-

uct’s overall superiority (Zeithaml 1988), and this vertical

differentiation across products allows for options to be ob-

jectively ranked (de Langhe et al. 2016; Tirole 1988).

Spiller and Belogolova (2017) recently found that individu-

als vary considerably in their quality assessment beliefs—

that is, their beliefs about the extent to which assessments

of a purchase are based on its objective quality. For in-

stance, when explaining their choice of one option over an-

other, some people are more likely to describe their chosen

option as being objectively better than the alternative and

to treat the superiority of their chosen option as a matter of

fact.

We propose that beyond differences across individuals

(Spiller and Belogolova 2017), quality assessment beliefs

may also differ across product domains—and in particular,

between experiential and material purchases. As an exam-

ple, people might believe that someone’s assessment of

their visit to a hot springs resort is based less on objective

quality than their assessment of a new hot tub. This is be-

cause hot springs visits (experiential purchases) are harder

to compare than hot tubs (material purchases; Carter and

Gilovich 2010), and “evaluations of quality usually take

place in a comparison context” (Zeithaml 1988, 5). Unlike

material products that are manufactured to be identical

such that all consumers should get the same thing “out of

the box,” experiences are necessarily unique to a particular

time and person (Eliashberg and Sawhney 1994). Indeed,

the same hot springs will produce quite different experien-

ces for two consumers depending on such factors as the

weather that day and who else is there, whereas two

consumers who purchase the same model of hot tub will

own identical products that deliver the same level of per-

formance. Additionally, attributes of material goods tend

to be more objective and quantifiable, which helps con-

sumers compare options along a continuum from worst to

best, whereas attributes of experiential goods tend to be

more subjective, which makes them harder to compare

across options (Holbrook and Hirschman 1982). For exam-

ple, hot tubs are judged along such features as the speed of

heating up, the pressure level and number of jets, and the

durability of material—all dimensions along which the

options can be objectively ranked and compared. In con-

trast, hot spring resorts are typically judged on such fea-

tures as d



ecor, view, and service, which are subjectively

evaluated. Altogether, because experiences are less compa-

rable across consumers, time, and options than material

possessions (Carter and Gilovich 2010; Rosenzweig and

Gilovich 2012), and quality judgments rely on such compa-

rability for a relative ranking of options (de Langhe et al.

2016; Spiller and Belogolova 2017; Zeithaml 1988), we

propose that people believe another’s assessment of an ex-

periential purchase is based less on objective quality than

another’s assessment of a material purchase.

We conducted a pilot study to explore our proposed link

between experiential versus material purchase type and

quality assessment beliefs. We first compiled a list of 87

purchases culled from 16 published papers that compared

experiential and material purchases. We then presented

these purchases (e.g., beach vacation package, concert

ticket, digital camera, stereo system) to participants

recruited on Amazon Mechanical Turk (MTurk) (N ¼ 263;

41% female, 2 unspecified; M

age

¼ 35.8). Each participant

saw a random subset of 10 purchases and rated the extent

to which each purchase was material or experiential (1 ¼

“primarily material,” 9 ¼ “primarily experiential”), as well

as the extent to which assessments of each purchase were a

matter of quality (1 ¼ “definitely not a matter of quality,”

9 ¼ “primarily a matter of quality”). We calculated the av-

erage material-experiential rating and the average quality

assessment beliefs rating for each purchase. The results

showed that the extent to which a purchase was viewed as

experiential (vs. material) was significantly and negatively

correlated with beliefs about its assessment as based on

quality (r ¼ –.27, p ¼ .01). That is, people viewed assess-

ments of experiential purchases as based less on objective

quality than material purchases. Similarly, when we cate-

gorized the purchases according to their treatment in the

prior papers (38 experiential purchases and 49 material

purchases), the results confirmed that people believed

assessments of experiential purchases to be based less on

objective quality (M ¼ 5.74, SD ¼ .93) than assessments of

material purchases (M ¼ 6.26, SD ¼ .97; t(85) ¼ 2.54, p ¼

.01, d ¼ .55). See web appendix A for the complete list of

papers and purchases, as well as the full survey and

analyses.

DAI, CHAN, AND MOGILNER 3

Downloaded from https://academic.oup.com/jcr/advance-article-abstract/doi/10.1093/jcr/ucz042/5567101 by UCLA Digital Collections Services user on 20 February 2020

These results offer preliminary evidence to suggest that

compared to material purchases, people believe that experi-

ential purchases involve assessments that are less based on

objective quality. Next, we theorize why such differences

would influence people’s reliance on consumer reviews,

such that people rely less on consumer reviews for experi-

ential purchases than for material purchases.

RELIANCE ON CONSUMER REVIEWS

People generally prefer advisors and are more willing to

use another’s behavior as a decision input when they be-

lieve the other’s judgment is objective rather than subjec-

tive (Gorenflo and Crano 1989; Olson, Ellis, and Zanna

1983; Spears, Ellemers, and Doosje 2009). For example,

males were more interested in knowing their peers’ ratings

of a female’s attractiveness when led to believe that beauty

is objective (vs. subjective; Olson et al. 1983). This ten-

dency to rely heavily on others’ objective judgments likely

translates into people’s tendency to rely more on consumer

reviews for purchases they believe to be assessed based on

objective quality.

Though people may search for and read reviews for a va-

riety of reasons, consumer reviews are useful to the extent

they help people predict what their own evaluations of an

option would be when (and if) consumed (Yaniv, Choshen-

Hillel, and Milyavsky 2011). Reviews that reflect the ob-

jective quality of an option are particularly predictive of

that option’s absolute value across consumers (Simonson

and Rosen 2014). Furthermore, when people believe there

to be less heterogeneity across people’s assessments of pur-

chase options, they expect others’ advice and reviews to be

more useful (Feick and Higie 1992; Price, Feick, and Higie

1989). Thus, people should be more likely to rely on con-

sumer reviews they perceive as based on objective quality.

Altogether, we hypothesize that compared to reviews for

material purchases, people believe reviews for experiential

purchases to be less based on the purchase’s objective

quality. We further hypothesize that this belief leads peo-

ple to rely less on consumer reviews for experiential pur-

chases than for material purchases.

OVERVIEW OF STUDIES

We tested these hypotheses in five studies. In study 1,

we analyzed archival data of over six million consumer

reviews posted on Amazon.com and found that people

were less likely to rate reviews as helpful for purchases

that were more experiential (vs. material). We then repli-

cated this finding in a series of experiments where we ma-

nipulated participants’ consideration of a more experiential

or material purchase by instructing participants to identify

an experiential or material purchase they planned to make

(study 3), or choose between given experiential product

options (cooking classes in study 2; ice cream shops in

studies 4 and 5) or material product options (espresso

machines in study 2; ice cream makers in studies 4 and 5).

In these studies, we measured participants’ reliance on con-

sumer reviews that they actually found online (study 3) or

that we adapted from actual online consumer reviews

(studies 2, 4, and 5). We measured review reliance in mul-

tiple ways: participants’ ratings of review helpfulness

(study 3), selection of the option with a more favorable re-

view (study 2), and likelihood of changing their purchase

decision after reading a slightly negative review (studies 4

and 5). We tested for our proposed mechanism of quality

assessment beliefs through both measurement (studies 3

and 4) and manipulation (study 5) and found support for its

role in the effect of experiential (vs. material) purchase

type on review reliance. In each study, the target sample

size was determined in advance of conducting the study,

and all data exclusions and manipulations are reported. All

measures are listed either in the article or in the web

appendixes.

STUDY 1: HELPFULNESS OF CONSUMER

REVIEWS ON AMAZON

Study 1 examined whether people find consumer

reviews posted on Amazon.com to be less helpful for expe-

riential (vs. material) purchases. Amazon is one of the

world’s leading sources for consumer reviews (Ante 2009;

Hong 2015). For each review, people shopping on Amazon

are asked, “Was this review helpful to you?” to which they

can voluntarily respond “Yes” or “No.” For reviews that

have received at least one vote, Amazon displays both the

number of “Yes” votes and the number of total votes. Prior

research shows that Amazon reviews rated as more helpful

have a stronger influence on shoppers’ purchase decisions

than reviews rated as less helpful (Chen, Dhanasobhon,

and Smith 2008). In study 1, we analyzed whether shoppers

are less likely to assign a “helpful vote” to Amazon

reviews for experiential purchases than for material

purchases.

Data

Our data comprised consumer reviews posted on

Amazon between January 31, 2008, and December 31,

2012 (see web appendix B for details about the data source:

McAuley and Leskovec 2013). For each review, we gath-

ered the product name, product category, review title, re-

view date, review text, star rating (1–5 stars), the number

of shoppers who responded either “Yes” or “No” to the

question asking whether the review was helpful (hereafter,

total votes), and the number of shoppers who responded

“Yes” to indicate that the review was helpful (hereafter,

helpful votes). Following past research that analyzed the

helpfulness of online reviews, we operationalized the

4 JOURNAL OF CONSUMER RESEARCH

Downloaded from https://academic.oup.com/jcr/advance-article-abstract/doi/10.1093/jcr/ucz042/5567101 by UCLA Digital Collections Services user on 20 February 2020

helpfulness of a review as the ratio of that review’s helpful

votes to its total votes in the main analysis, thus excluding

reviews that received zero total votes from the analysis

(Forman et al. 2008; Mudambi and Schuff 2010; Yin et al.

2014). The final data included 6,508,574 reviews written

about 997,845 products.

Measures

Review Helpfulness. The dependent variable equaled

the ratio of the number of helpful votes a review received

to the number of total votes it received (Forman et al.

2008; Mudambi and Schuff 2010; Yin et al. 2014). Helpful

votes exceeded total votes for 31 of the 6,508,574 reviews.

We assigned a ratio of 1 to these extremely rare cases of

data error, but the results were robust when we excluded

these reviews from the analysis.

Experiential (vs. Material) Rating. The primary inde-

pendent variable was the extent to which a given product

was experiential or material. Given that the data included

approximately one million different products, we assessed

the experiential (vs. material) rating of each product based

on its product category. To obtain these ratings, we

recruited 100 participants from MTurk to complete a prod-

uct survey in exchange for $.50 (see web appendix B for

the complete survey). We instructed participants to imag-

ine they were shopping on Amazon and presented them

with a list of 26 product categories (see table 1 for catego-

ries). These categories reflected Amazon’s 26 top-level

category labels (e.g., books, music, shoes), except for cases

that required additional description for clarity (e.g., we de-

scribed “Amazon instant videos” as “videos for stream-

ing”). After defining material and experiential purchases

(i.e., “material purchases provide something that a person

can keep in his/her possession, and experiential purchases

provide something that a person can do”; adapted from

Van Boven and Gilovich 2003), we asked participants to

rate the extent to which products in each category were

material or experiential (1 ¼ “purely material,” 9 ¼

“purely experiential”). Since interrater reliability was high

(intraclass correlation coefficient ¼ .98), we averaged par-

ticipants’ responses to form an experiential (vs. material)

rating for each product category.

Other Measures. We accounted for a number of other

factors that also might influence review helpfulness. First,

given that review helpfulness differs between hedonic

products (those purchased primarily for pleasure and fun)

and utilitarian products (those purchased primarily out of

necessity and for practical functions; Chu, Roh, and Park

2015; Moore 2015; Sen and Lerman 2007), we also asked

the product survey participants to rate the extent to which

products in each category were utilitarian or hedonic (1 ¼

“purely utilitarian,” 9 ¼ “purely hedonic”). Participants’

responses were averaged to form a hedonic rating for each

product category (intraclass correlation coefficient ¼ .99).

Second, because people might care more about their ex-

periential purchases than their material purchases (Nicolao

et al. 2009), which could influence their reliance on

reviews, we also asked the product survey participants to

rate how much they would care about their purchase de-

cision when shopping for a product in each of the 26

categories (1 ¼ “not at all,” 7 ¼ “very much”).

Participants’ responses were averaged to form a caring

rating for each product category (intraclass correlation

coefficient ¼ .94).

Following past research (Mudambi and Schuff 2010;

Yin et al. 2014, 2017; Forman et al. 2008), we compiled a

number of other review characteristics that could influence

review helpfulness. Specifically, for each review, we in-

cluded the number of words in the text (review length), the

number of words in the title (title length), the star rating

given (star rating), the number of days between its posting

and the final date of data collection, March 4, 2013 (review

age), and the total number of reviews the product in ques-

tion had (including reviews with zero total votes and those

with at least one total vote; review availability).

Table 1, panel A, displays descriptive statistics for each

product category in terms of experiential (vs. material) rat-

ing, review helpfulness, and other aforementioned meas-

ures. Table 1, panel B, displays the summary statistics of

and correlations between these measures.

Results

As a first pass, we conducted a category-level analysis

by calculating the average proportion of helpful votes

across all reviews within a given product category

(figure 1).

In general, shoppers voted reviews as helpful: on aver-

age across all purchase categories, Amazon reviews were

identified as helpful by 69% of shoppers who voted. The

highest percentage of helpful votes was in the jewelry cate-

gory (86%), and the lowest percentage was in the movies

and TV shows category (59%), which was still the majority

of shoppers who voted. More importantly, consistent with

our hypothesis, the experiential rating of a product cate-

gory was negatively correlated with the percentage of help-

ful votes that reviews in this product category received

(r ¼ –.69, p < .0001; N ¼ 26). This suggests that the more

experiential the product category, the less likely shoppers

were to view the consumer reviews as helpful.

For a more precise analysis, we next turned to Ordinary

Least Squares (OLS) regressions at the review level, with

review helpfulness as the dependent variable. Similar to

prior work that analyzed the helpfulness of Amazon

DAI, CHAN, AND MOGILNER 5

Downloaded from https://academic.oup.com/jcr/advance-article-abstract/doi/10.1093/jcr/ucz042/5567101 by UCLA Digital Collections Services user on 20 February 2020

TABLE 1

AMAZON CONSUMER REVIEW DATA DESCRIPTIVE STATISTICS (STUDY 1)

PANEL A: SUMMARY STATISTICS ACROSS PRODUCT CATEGORIES

Product category

(descriptions used

in our survey)

Experiential

rating

(1–9)

Review

helpfulness

(proportion

of helpful

votes)

Hedonic

rating

(1–9)

Caring

rating

(1–7)

Review

length

(words)

Title

length

(words)

Star

rating

(1–5)

Review

age

(in days)

Review

availability

(number of

reviews per

product)

Number

reviews

Number

products

Shoes 2.57 77.16% 3.78 5.29 76.89 3.82 4.01 899.03 15.49 103,048 16,131

Watches 2.58 79.09% 4.75 4.48 95.20 4.00 3.94 1080.63 10.33 16,276 3,046

Ofﬁce products 2.62 80.19% 2.68 4.00 91.16 4.26 3.81 964.99 10.34 36,736 6,503

Jewelry 2.72 85.75% 7.34 4.57 59.46 3.37 4.01 1180.35 5.06 15,291 5,710

Electronics 2.84 75.61% 4.08 5.38 112.62 4.48 3.68 1169.31 17.45 286,203 33,817

Home and kitchen

products (house-

wares,

furnishings,

accessories)

2.89 82.59% 3.01 4.82 101.34 4.18 3.76 992.87 15.69 313,685 35,210

Automotive (parts,

accessories,

tools, or

equipment)

2.95 78.79% 2.61 4.61 87.13 4.18 3.86 838.97 6.40 61,218 17,563

Pet supplies 3.02 83.35% 2.89 4.51 108.38 4.38 3.79 867.88 16.58 76,192 9,653

Patio, lawn, and gar-

den products

3.15 82.57% 3.95 4.13 99.60 4.14 3.82 931.41 10.74 48,355 7,862

Baby products 3.26 78.64% 2.61 4.33 118.95 4.44 3.67 1040.25 30.44 32,944 2,950

Tools and home

improvement

3.73 79.01% 2.61 4.66 101.67 4.26 3.82 930.73 8.94 113,799 21,707

Scientiﬁc, lab, and

industrial supplies

3.75 73.53% 2.87 4.06 54.00 4.55 4.44 703.66 5.94 42,471 11,757

Beauty products 3.88 79.27% 6.15 4.16 86.96 3.99 3.91 887.32 11.12 87,684 14,853

Clothing and

accessories

3.89 79.39% 5.72 6.09 78.65 3.93 3.88 892.01 12.40 156,288 27,315

Health and personal

care

4.66 78.97% 3.50 5.28 93.38 4.11 3.88 872.29 12.69 159,564 21,346

Grocery and gour-

met foods

4.74 78.87% 3.92 5.46 81.06 4.14 4.01 981.17 8.02 56,084 12,112

Books 4.95 67.34% 5.72 5.14 164.85 4.76 3.99 982.65 6.97 2,220,093 374,686

Musical instruments 4.95 79.34% 6.27 4.49 105.76 4.23 4.03 889.79 9.90 30,174 7,302

Toys and board

games

5.05 80.51% 7.32 4.32 91.21 4.25 3.87 1049.11 8.77 88,300 20,525

Software 5.18 72.79% 4.88 5.30 116.68 4.64 3.20 1372.30 6.87 11,050 2,407

Arts, crafts, and

sewing

5.20 84.23% 5.80 3.86 88.93 3.93 3.97 790.51 7.19 9,866 2,080

Sports and outdoors 5.78 79.29% 5.82 4.52 100.87 4.09 3.92 929.90 10.38 140,771 24,015

Video games 5.93 60.15% 7.90 5.28 147.97 4.65 3.59 1237.53 12.99 56,681 8,532

Music 6.64 69.78% 7.38 5.08 156.15 4.78 4.23 1144.76 4.99 503,410 170,374

Movies and TV

shows

6.70 58.91% 7.87 5.36 153.68 4.76 3.92 975.10 20.91 1,612,171 122,120

Videos for

streaming

7.12 59.66% 7.90 4.92 161.61 4.83 3.70 915.30 19.68 230,220 18,269

All 4.26 68.58% 4.97 4.77 141.18 4.60 3.93 989.15 11.22 6,508,574 997,845

NOTE.—In Table 1, Panel A, for experiential, hedonic, and caring ratings that were originally collected at the category level, the last row reports the mean value

across 26 product categories. For review availability that was originally collected at the product level, the last row reports the mean value across all products. For

variables that were originally collected at the review level (including review helpfulness, review length, title length, star rating, and review age), the last row reports

the mean value across all reviews in the data. For the number of reviews and products, the last row reports the total number of reviews and products in the data.

In Table 1, Panel A, product categories are ordered by experiential ratings (1 ¼ “purely material,” 9 ¼ “purely experiential”).

6 JOURNAL OF CONSUMER RESEARCH

Downloaded from https://academic.oup.com/jcr/advance-article-abstract/doi/10.1093/jcr/ucz042/5567101 by UCLA Digital Collections Services user on 20 February 2020

reviews (Forman et al. 2008), we relied on the following

OLS regression specification:

Review helpfulness

ijk

¼ a

þ b experiential rating

ðÞ

þ X

ijk

þ e

ijk

where i indexes the review, j indexes the product, k indexes

the product category, X

ijk

is the vector of control variables,

and e

ijk

is the error term. Since more than 40% of the prod-

ucts had more than one review in our final data and the er-

ror terms are not independent among consumer reviews for

the same product, we clustered standard errors at the prod-

uct level.

We began with a regression that did not include any con-

trol variables and simply used experiential rating to predict

review helpfulness. The results of this basic model sup-

ported our prediction and are reported in model 1 in

table 2. Reviews for more experiential purchases were less

likely to be rated as helpful than reviews for less

experiential purchases (B ¼ –.0480, SE ¼ .0004, p <

.001). More specifically, a one-point increase in the experi-

ential rating on the nine-point Likert scale was associated

with an average 4.8 percentage-point decrease in the pro-

portion of people who found a review helpful.

Next, we ran a full model including the control varia-

bles. Again, consistent with our hypothesis, the relation-

ship between experiential rating and review helpfulness

remained negative and significant (B ¼ –.0368, SE ¼

.0009, p < .001; model 2 in table 2). Specifically, a one-

point increase in the experiential rating was associated

with an average 3.7 percentage-point decrease in the pro-

portion of people who found a review helpful. Though the

various control variables could not fully explain the effect

of experiential (vs. material) purchases on review helpful-

ness, the regression results (model 2 in table 2) did support

past research in showing that reviews for more hedonic

products were viewed as less helpful (B ¼ –.0131, SE ¼

.0007, p < .001; Sen and Lerman 2007).

Robustness Checks. The results of the OLS regressions

(models 1 and 2 in table 2) remained unchanged in terms

of magnitude and statistical significance irrespective of

whether we (1) clustered standard errors at the product cat-

egory level, (2) estimated standard errors without cluster-

ing, (3) omitted the 31 reviews for which helpful votes

exceeded total votes, or (4) used Tobit regression models

(Mudambi and Schuff 2010; Yin et al. 2014).

We also modeled helpful votes as an alternative depen-

dent variable and controlled for total votes (Yin et al.

2017) using various regression specifications. Because

most reviews in our sample received few helpful votes and

a small number of reviews received thousands of helpful

votes, the alternative dependent measure, helpful votes

(mean ¼ 3.05, variance ¼ 538.28), exhibited overdisper-

sion (overdispersion parameter ¼ 1.35, p < .0001 for the

log-likelihood ratio test of the null hypothesis that the over-

dispersion parameter equals zero). Therefore, we ran stan-

dard negative binomial regression models (instead of

FIGURE 1

CATEGORY-AVERAGE PROPORTION OF HELPFUL VOTES AS

A FUNCTION OF PRODUCTS’ EXPERIENTIAL RATINGS

(STUDY 1)

Jewelry

Electronics

Home and kitchen products

Beauty products

Books

Video games

Music

Movies and TV

shows

Videos for

streaming

50%

60%

70%

80%

90%

2345678

Category–average proportion of

helpful votes

eriential ratin

roduct cate

PANEL B: SUMMARY STATISTICS OF AND CORRELATIONS AMONG MEASURES

Mean SD 1 2 3 4 5 6 7 8

1. Experiential rating 5.22 1.36

2. Review helpfulness 68.58% 39.26% –.1657*

3. Hedonic rating 6.02 1.65 .9184* –.1656*

4. Caring rating 5.12 .34 .3052* –.0967* .3685*

5. Review length (words) 141.18 187.65 .0998* .1259* .0951* .0503*

6. Title length (words) 4.60 2.96 .0633* .0730* .0579* .0277* .2814*

7. Star rating 3.93 1.42 .0321* .2681* .0260* .0012* .0016* –.0246*

8. Review age (in days) 989.15 536.75 .0124* .0352* .0280* .0258* .0326* .0284* .0413*

9. Review availability 163.40 375.58 .0044* –.0667* .0179* .0681* –.0035* –.0033* –.0244* –.0612*

NOTE.— Table 1, Panel B, reports the summary statistics of and raw correlation among the independent measure, dependent measure, and covariates at the

review level (without clustering standard errors). *p < .05.

TABLE 1 (CONTINUED)

DAI, CHAN, AND MOGILNER 7

Downloaded from https://academic.oup.com/jcr/advance-article-abstract/doi/10.1093/jcr/ucz042/5567101 by UCLA Digital Collections Services user on 20 February 2020

Poisson regressions). To reduce computational complexity,

we excluded consumer reviews whose total votes were

more than three standard deviations above the mean (i.e.,

more than 80 votes), which accounted for .3% of all

reviews in the final dataset.

Models 3 and 4 in table 2 re-

port the results of the negative binomial regressions, indi-

cating that reviews for more experiential purchases

received fewer helpful votes. Specifically, a one-point in-

crease in the experiential rating on the nine-point Likert

scale was associated with a decrease in the number of help-

ful votes by 6.33% (based on model 3; i.e., [e

–.0654

–1] 

100%) or 2.90% (based on model 4; i.e., [e

–.0294

–1] 

100%). We observed the same patterns when we used (1)

zero-inflated negative binomial regressions, (2) Tobit mod-

els with the left limit of the dependent variable set equal to

zero, and (3) OLS regressions with either helpful votes or

the log-transformed helpful votes as the dependent variable

(web appendix B).

Discussion

Study 1 examined helpful votes provided by thousands

of actual Amazon shoppers for millions of real consumer

reviews. Although reviews were generally identified as

helpful across purchase categories, consistent with our hy-

pothesis, people were less likely to identify consumer

reviews as helpful for products deemed more experiential.

The results were robust to different model specifications

and data inclusion criteria.

Despite this study’s tremendous external validity, it has

several limitations that we sought to address in the subse-

quent experiments. First, the context of Amazon reviews

excludes the highly experiential purchases (e.g., restaurant

meals, event tickets, and vacations) that people typically

think of and that the literature comparing experiential and

material purchases has often examined. Even though there

was high variation in experiential ratings across Amazon’s

26 product categories (ranging from 2.57 to 7.12), 14 of

the categories were rated as clearly material (significantly

below the scale midpoint), whereas only five of the cate-

gories were rated as clearly experiential (significantly

above the scale midpoint). Thus, to include experiences

such as dining out and vacations, all of the remaining stud-

ies examined prototypical experiential and material pur-

chases. In particular, study 3 asked participants about an

experiential (or material) purchase they planned to make,

which allowed us to test our hypotheses across a broad

range of purchases—including the highly experiential pur-

chases that fill popular websites (e.g., Yelp and

TripAdvisor).

Second, because it is impossible to know the number of

shoppers who read a review but did not provide a helpful

or unhelpful vote, we could not assess the true proportion

of readers who found a review helpful conditional on read-

ing the review. Notably, these results therefore reflect the

relationship between how experiential a purchase was and

the likelihood that shoppers gave a helpful vote conditional

on rating a review (as opposed to reading a review). In the

remaining studies, we use a variety of dependent measures

that more precisely capture the extent to which shoppers

find the reviews they read helpful and rely on the reviews

in their decision-making.

Another limitation of this archival field study is that de-

spite controlling for a number of alternative accounts, due

to the correlational nature of this data we were not able to

establish a causal relationship between purchase type and

review helpfulness. We thus conducted the subsequent

experiments to test the causal effect of experiential (vs.

material) purchases on people’s reliance on consumer

reviews.

TABLE 2

REVIEW HELPFULNESS OF AMAZON REVIEWS AS A

FUNCTION OF PRODUCTS’ EXPERIENTIAL RATINGS

(STUDY 1)

Dependent variable

Review helpfulness

measured by

proportion of helpful

votes (0%–100%)

Review

helpfulness

measured

by number of

helpful votes

Predictor variables

Model 1 Model 2 Model 3

Model 4

Experiential rating –.0480

–.0368

–.0654

–.0294

(.0004) (.0009) (.0009) (.0024)

Hedonic rating –.0131

–.0401

(.0007) (.0018)

Caring rating –.0488

–.0739

(.0014) (.0037)

Review length .0003

.0006

(2.06e-06) (4.80e-06)

Title length .0073

.0169

(.0001) (.0001)

Star rating .0756

.0599

(.0008) (.0022)

Review age 1.40e-05

.0002

(1.28e-06) (3.56e-06)

Review availability –5.63e-05

–5.92e-05

(1.29e-05) (4.03e-05)

Number of total

votes

.1019

.0986

(.0004) (.0004)

Observations 6,508,574 6,508,574 6,487,944 6,487,944

or pseudo R

.03 .14 .16 .18

Model speciﬁcation OLS Negative binomial

regression

These models exclude reviews whose total votes are more than three

standard deviations above the mean (because including these reviews with

rarely large total votes would cause the models to fail to converge).

Indicates signiﬁcance at the .1% level. Standard errors are clustered at

the product level and are reported in parentheses.

1 When all reviews in our final dataset were included, the negative

binomial regression models were unable to converge. A careful exami-

nation suggested that the rare large values of helpful votes (maximum

¼ 32,208, mean ¼ 3.05) and total votes (maximum ¼ 32,506, mean ¼

4.75) were responsible.

8 JOURNAL OF CONSUMER RESEARCH

Downloaded from https://academic.oup.com/jcr/advance-article-abstract/doi/10.1093/jcr/ucz042/5567101 by UCLA Digital Collections Services user on 20 February 2020

STUDY 2: INFLUENCE OF CONSUMER

REVIEWS ON CHOICE

Study 2 experimentally tested whether people rely less

on consumer reviews for experiential purchases than for

material purchases. We assessed reliance on consumer

reviews by testing the influence of reviews among partici-

pants deciding between two experiential or two material

purchase options. We randomly varied which option within

each choice pair received a positive or negative review,

and predicted that participants deciding between the expe-

riential options would be less swayed to choose the posi-

tively reviewed option compared to participants deciding

between the material options. To make this a real choice,

we entered participants into a drawing, and winners re-

ceived their chosen option.

Method

Participants. We recruited 212 university students to

complete this study. Since this research focuses on people

who are in the process of choosing what to buy, nine par-

ticipants who had previously consumed the products were

excluded from the analysis. Our analysis focuses on the

remaining 203 participants (50% female, M

age

¼ 20.4).

Purchase Type Manipulation. In this between-subjects

study, participants were presented with either two options

of cooking classes (experiential condition) or two options

of espresso machines (material condition) and asked to

choose which they would prefer. They were informed that

one in every 100 participants would be randomly selected

to receive their chosen option. All options were valued at

$95–$100.

These purchases were selected based on a pretest

(N ¼ 172) showing that cooking classes and espresso

machines differed in how experiential (vs. material) they

were (1 ¼ “purely material,” 9 ¼ “purely experiential”;

cooking_class

¼ 6.44, SD ¼ 1.89 vs. M

espresso_machine

4.42, SD ¼ 2.30; t(102) ¼ 4.89, p < .0001), but not in par-

ticipants’ caring, desire, or knowledge of these purchases

(all ps > .33).

Furthermore, study 1 found that material-

experiential ratings were positively correlated with

utilitarian-hedonic ratings (with experiences being viewed

as more hedonic) and that review helpfulness differed be-

tween hedonic and utilitarian products (as in past research;

Chu et al. 2015; Moore 2015; Sen and Lerman 2007).

Given these observations, we addressed the possibility that

the hedonic (vs. utilitarian) nature of the purchases might

explain the effect of experiential (vs. material) purchase

type by selecting a pair of purchases in which the

experiential purchase was not more hedonic than the mate-

rial purchase. A separate pretest (N ¼ 80) showed that a

cooking class was, in fact, viewed as less hedonic than an

espresso machine (1 ¼ “purely utilitarian,” 9 ¼ “purely

hedonic”; M

cooking_class

¼ 4.82, SD ¼ 1.47 vs.

espresso_machine

¼ 6.41, SD ¼ 2.12; t(78) ¼ 3.89, p ¼

.0002).

In the main study, for each of the two options, we pre-

sented participants with a picture and product description

(46–48 words, adapted from its online product information),

along with one consumer review. The two options within

each choice pair were presented side-by-side, with the order

randomized. See appendix A for the study stimuli.

Consumer Review Manipulation. Within each choice

pair, we randomly assigned a positive 5-star review to one

option and a more negative 3-star review to the other op-

tion, thereby counterbalancing which option received the

positive or negative review. These reviews were adapted

from real 5- and 3-star online consumer reviews for cook-

ing classes on Yelp and for espresso machines on Amazon.

We ensured that all reviews had the same length (83–85

words), and all reviews of the same valence had the same

structure (see appendix A). For example, positive 5-star

reviews in each condition had eight sentences, beginning

with the sentence “This is the best cooking class I have tak-

en [espresso machine I have owned],” and concluding with

the sentence “I would never have thought I could do it so

easily!”

A pretest (N ¼ 120) confirmed that for both the cooking

classes and espresso machines, the positive reviews were

viewed as significantly more positive than the negative

reviews (–3 ¼ “very negative,” 3 ¼ “very positive”; M

5-star

cooking class

¼ 2.62, SD ¼ .93 vs. M

3-star cooking class

¼ .38,

SD ¼ 1.07; t(60) ¼ 16.88, p < .0001; M

5-star espresso machine

¼ 2.49, SD ¼ 1.15 vs. M

3-star espresso machine

¼ .37, SD ¼

1.16; t(58) ¼ 13.90, p < .0001). Also, within each pair of

options, the reviews were seen as significantly favoring the

option with the positive review over the option with the

negative review (–3 ¼ “the reviews definitely favor Option

A,” 3 ¼ “the reviews definitively favor Option B”;

cooking class

¼ 2.39, SD ¼ 1.05 vs. the scale midpoint of

0; t(60) ¼ 17.75, p < .0001; M

espresso machine

¼ 2.25, SD ¼

1.31 vs. the scale midpoint of 0; t(58) ¼ 13.24, p < .0001).

Importantly, there were no significant differences between

the cooking classes and espresso machines in terms of the

positivity of the positive reviews, the negativity of the neg-

ative reviews, or the extent to which the reviews favored

the option with the positive review over the option with the

negative review (all ps > .49). See web appendix C for the

results of all study 2 pretests and the procedure for generat-

ing consumer reviews.

Reliance on Consumer Reviews. Participants chose one

option from the two presented to them. The dependent

measure was whether a participant chose the option with

2 Pretest participants were presented with three purchases randomly

selected from 10 purchases.

DAI, CHAN, AND MOGILNER 9

Downloaded from https://academic.oup.com/jcr/advance-article-abstract/doi/10.1093/jcr/ucz042/5567101 by UCLA Digital Collections Services user on 20 February 2020

the positive review over the option with the negative re-

view. This measure thereby assessed whether participants

relied on the consumer reviews to make their choice.

Manipulation Check. Participants rated the extent to

which a cooking class [an espresso machine] is a material

or experiential purchase (1 ¼ “purely material,” 9 ¼

“purely experiential”).

Other Measures. Participants indicated whether they

had previously heard of or visited either of the cooking

schools (or had used either of the brands of espresso ma-

chine), and rated how knowledgeable they were about

cooking classes (or espresso machines; 1 ¼ “not at all

knowledgeable,” 7 ¼ “very knowledgeable”).

Results and Discussion

Manipulation Check. The manipulation check con-

firmed that a cooking class is viewed as more experiential

(M ¼ 6.08, SD ¼ 1.75) than an espresso machine

(M ¼ 4.85, SD ¼ 2.02; t(201) ¼ 4.62, p < .0001, d ¼ .65).

Reliance on Consumer Reviews. Participants in the ex-

periential condition were less likely to choose the option

with the positive review (66.34%) than those in the mate-

rial condition (79.41%; v

(1) ¼ 4.39, p ¼ .036

). This sug-

gests that though many participants relied on the consumer

reviews when deciding on an experiential purchase, even

more did when deciding on a material purchase. There was

not a significant difference in participants’ knowledge about

these purchases (M

experiential

¼ 2.33, SD ¼ 1.33 vs. M

material

¼ 2.38, SD ¼ 1.50; t(201) ¼ .28, p ¼ .78, d ¼ .04).

One potential alternative explanation for this study’s

results is that the product information revealed more differ-

ences between the two cooking class options than between

the two espresso machine options. With more product in-

formation to distinguish between the experiential options,

people would not need to rely as much on the consumer

reviews to make their selection in the experiential condi-

tion as in the material condition. However, this did not

seem to be the case based on a pretest we conducted. In a

within-subjects pretest (N ¼ 30), we presented participants

with the same names, pictures, and product descriptions

(without a consumer review) for each pair of options in

random order, and asked participants to indicate which of

the two options they would be more likely to choose. We

found that the relative preferences for the two options did

not differ between the two purchase types (M

experiential

4.80, SD ¼ 1.69 vs. M

material

¼ 4.77, SD ¼ 1.87; t(29) ¼

.07, p ¼ .94).

Thus, the observed difference in reliance on

consumer reviews in the main study was unlikely driven by

variations in the provided product information.

Study 2 leveraged random assignment through an exper-

imental design and corroborated the results of study 1’s

field data. Using a behavioral measure for reliance on con-

sumer reviews and an incentive-compatible design, study 2

provided further evidence that people rely less on con-

sumer reviews when making an experiential purchase than

when making a material purchase.

We conducted two additional experiments that manipu-

lated participants’ focus on the experiential or material

aspects of the same purchase (a sleeping bag in web appen-

dix D and a mattress in web appendix E). Similar to Carter

and Gilovich (2010; study 6), this approach is based on the

insight that many purchases have both experiential and ma-

terial attributes (e.g., a mattress is a material possession

that offers the experience of a good night’s sleep). The

results indicated that participants relied less on consumer

reviews in the experiential conditions than in the material

conditions. With all aspects of the purchase held constant

other than its experiential or material nature, these studies

provide a conservative conceptual replication of the effect

observed in study 2.

The next study again tests for the effect of experiential

(vs. material) purchase type on review reliance, but in the

context of real consumer reviews for purchases that partici-

pants actually plan to make. Further, the next study

explores the psychological underpinnings for the effect.

STUDY 3: THE ROLE OF QUALITY

ASSESSMENT BELIEFS

Study 3 sought to conceptually replicate the effect ob-

served in study 2 across a broad array of actual planned

purchases by asking participants to describe either an ex-

periential or material purchase they intended to make in

the coming year. To further enhance realism, participants

then searched online for and read real consumer reviews

about their purchase, and reported the usefulness of these

reviews for their purchase decision. Finally, this study

tested the proposed mechanism by asking participants to

report the extent to which the reviews reflected the previ-

ous consumers’ assessments of the purchase’s objective

quality.

3 We conducted a logit regression to predict whether participants

chose the option with a positive review as a function of the experien-

tial (vs. material) manipulation. The binary indicator for the experien-

tial condition had an odds ratio of 0.51 (SE ¼ 0.16, p ¼ 0.038), which

can be translated into a Cohen’s d of –.37.

4 This pretest (described in detail in web appendix C) additionally

confirmed that the pair of cooking classes and the pair of espresso

machines did not significantly differ in terms of (1) how much partici-

pants desired each pair of options (p ¼ .57), (2) how much participants

cared about choosing between the two options within the pair (p ¼

.82), and (3) how familiar participants were with each option (ps for

all pairwise comparisons > .14).

10 JOURNAL OF CONSUMER RESEARCH

Downloaded from https://academic.oup.com/jcr/advance-article-abstract/doi/10.1093/jcr/ucz042/5567101 by UCLA Digital Collections Services user on 20 February 2020

Method

Participants. A total of 301 participants (54% female;

age

¼ 36.7, 2 unspecified) recruited through MTurk com-

pleted this study in exchange for $1.00.

Purchase Type Manipulation. In this between-subjects

study, participants were randomly assigned to specify ei-

ther an experiential or material purchase that cost at least

$50, which they planned to make in the coming year.

Participants in the experiential condition were instructed to

think of a purchase that “involves spending money with the

primary intention of acquiring a life experience—an event

or series of events that you personally will encounter or

live through.” Participants in the material condition were

instructed to think of a purchase that “involves spending

money with the primary intention of acquiring a material

possession—a tangible object that you obtain and keep in

your possession” (adapted from Van Boven and Gilovich

2003). To control for the possibility that experiential pur-

chases may be more hedonic or have fewer reviews avail-

able than material purchases, we additionally instructed

participants in both conditions to list a fun and enjoyable

purchase that had online consumer reviews available.

Reliance on Consumer Reviews. Participants were then

instructed to search for and read five online reviews written

by other consumers about the purchase they had specified.

They were told that they could look for consumer reviews

anywhere online and were asked to copy the five reviews

they read into the survey. Importantly, participants spent a

similar amount of time searching for and reading reviews

between the experiential (M ¼ 245.49 seconds, SD ¼

148.06) and material conditions (M ¼ 267.28 seconds, SD ¼

184.21; t(299) ¼ 1.11, p ¼ .27, d ¼ .13), suggesting that it

was not harder for participants to find reviews for one pur-

chase type than the other. For the consumer reviews they

read, participants reported how helpful the reviews were

(1 ¼ “not at all,” 7 ¼ “extremely”), how useful the reviews

were (1 ¼ “not at all,” 7 ¼ “extremely”), and how much

they would rely on these consumer reviews for their pur-

chase decision (1 ¼ “not at all,” 7 ¼ “very much”). These

three items were averaged to serve as the primary dependent

variable: reliance on consumer reviews (a ¼ .93).

Quality Assessment Beliefs. To test the proposed mech-

anism, on the next page, we asked participants to indicate

the extent to which the reviews they read reflected other

consumers’ objective assessments of the purchase’s quality

(1 ¼ “not at all,” 7 ¼ “a great deal”), which was adapted

from Spiller and Belogolova (2017).

Manipulation Check. Participants rated the extent to

which their planned purchase was material or experiential

(1 ¼ “purely material,” 9 ¼ “purely experiential”).

Other Measures. Participants listed the website(s) on

which they found the consumer reviews. For controls, we

also asked participants to indicate the number of different

options their five reviews covered, as well as to rate how

much they cared about their purchase decision, how impor-

tant the purchase was to them, and how engaged they were

in this purchase decision (1 ¼ “not at all,” 7 ¼ “very

much”). These last three items were averaged to create a

measure of purchase importance (a ¼ .82). Finally, partici-

pants rated how knowledgeable they were about the pur-

chase (1 ¼ “not at all,” 7 ¼

“very much”) and indicated

how much money they would spend on the purchase. See

web appendix F for the complete measures.

Results

Manipulation Check. Participants listed a wide range

of hedonic experiential purchases (e.g., dinner at a restau-

rant, vacations, and event tickets) and hedonic material

purchases (e.g., home accessories, fun clothing items, and

electronic gadgets). Participants in the experiential condi-

tion rated their planned purchase as more experiential

(M ¼ 7.81, SD ¼ 1.68) than participants in the material

condition (M ¼ 3.36, SD ¼ 2.29; t(299) ¼ 18.86,

p < .0001, d ¼ 2.19).

Reliance on Consumer Reviews. Participants largely

found reviews on popular websites such as Yelp, TripAdvisor,

Amazon, and Best Buy. On average, participants rated the

reviews they read to be somewhat useful (M ¼ 5.64, p <

.0001 vs. the midpoint of 4); however, they found the con-

sumer reviews for experiential purchases (M ¼ 5.44, SD ¼

1.39) to be less useful than those for material purchases (M ¼

5.80, SD ¼ 1.15; t(299) ¼ 2.49, p ¼ .01, d ¼ .29).

Mediation by Quality Assessment Beliefs. To test the

proposed mechanism, we examined the role of quality as-

sessment beliefs. As predicted, compared to material pur-

chases (M ¼ 5.28, SD ¼ 1.40), participants perceived that

reviews of an experiential purchase reflected quality

assessments to a lesser degree (M ¼ 4.86, SD ¼ 1.41;

t(299) ¼ 2.57, p ¼ .01, d ¼ .30). Moreover, a 5,000-sample

bootstrap analysis (model 4 in Hayes 2013) estimated an

indirect effect of purchase type on reliance on reviews via

quality assessment beliefs as –.13 (SE ¼ .06), and the 95%

bias-corrected confidence interval (CI) of the indirect ef-

fect did not include zero ([–.26, –.04]). Thus, quality as-

sessment beliefs mediated the influence of purchase type

on review reliance (figure 2).

Other Measures. The results indicated no significant

differences between experiential and material purchases in

the number of options discussed by the reviews participants

read (M

experiential

¼ 1.49, SD ¼ 1.05 vs. M

material

¼ 1.46, SD

¼ 1.09; t(292) ¼ .23, p ¼ .81, d ¼ .03), purchase importance

experiential

¼ 6.03, SD ¼ .96 vs. M

material

¼ 5.90, SD ¼

.97; t(299) ¼ 1.20, p ¼ .23, d ¼ .14), purchase knowledge

experiential

¼ 5.41, SD ¼ 1.27 vs. M

material

¼ 5.43, SD ¼

1.24; t(299) ¼ .18, p ¼ .86, d ¼ .02), or log-transformed

DAI, CHAN, AND MOGILNER 11

Downloaded from https://academic.oup.com/jcr/advance-article-abstract/doi/10.1093/jcr/ucz042/5567101 by UCLA Digital Collections Services user on 20 February 2020

expected cost (M

experiential

¼ 5.33, SD ¼ 1.21 vs. M

material

5.29, SD ¼ 1.04; t(298) ¼ .26, p ¼ .80, d ¼ .03).

Furthermore, when we included all of these measures in a

multi-mediator model (model 4 in Hayes 2013) along with

quality assessment beliefs, quality assessment beliefs was

the only significant mediator (indirect effect ¼ –.11, SE ¼

.05, 95% CI ¼ [–.24, –.03]).

Discussion

Amongst an array of experiential and material pur-

chases that had consumer reviews readily available, peo-

ple found the reviews they read to be less useful for

experiential purchases than for material purchases.

Indeed, people viewed assessments of experiential pur-

chases to be less a matter of quality than material pur-

chases, which drove the effect. This study explored a

number of potential alternative explanations, none of

which gained supportive evidence. Overall, study 3’s

results provide initial evidence that quality assessment

beliefs play a unique role in the influence of purchase

type on the extent to which people rely on consumer

reviews. Though having participants search for and read

online consumer reviews for purchases they were actually

planning to make provided external validity, we sought to

conceptually replicate these findings using more tightly

controlled stimuli in the next study.

STUDY 4: CONTROLLING REVIEW

CONTENT, ANOTHER TEST OF

MECHANISM

The objective of study 4 was to provide further evidence

that quality assessment beliefs are responsible for the dif-

ference in review reliance between experiential and mate-

rial purchases. Study 4 built on the previous studies in two

ways. First, like study 2, study 4 involved an incentive-

compatible design, but used a new behavioral measure for

reliance on consumer reviews: whether participants

changed their mind after reading a negative review about

their initial choice. Second, in addition to using a pair of

experiential and material purchases that were in the same

hedonic consumption domain (ice cream), study 4 pre-

sented all participants with a virtually identical consumer

review. Not only did this ensure a high degree of experi-

mental control, but it allowed us to test whether people

perceive assessments written in consumer reviews for ex-

periential purchases as less reflective of the purchase’s ob-

jective quality than those for material purchases and,

consequently, rely less on reviews for experiential

purchases.

Method

Participants. We recruited 238 participants from a uni-

versity’s subject pool that included students and commu-

nity members. Given that this research focuses on people

who are in the process of choosing what to buy, as in study

2, we excluded from the analysis 19 participants who had

previously consumed the products. In addition, one partici-

pant who reported not being able to view the stimuli in the

online survey was excluded from the analysis. The final

sample included 218 participants (67% female, 3 unspeci-

fied; M

age

¼ 23.6, 1 unspecified).

Purchase Type Manipulation. In this between-subjects

study, participants were presented with either two options

of nearby ice cream shops (experiential condition) or two

options of ice cream makers that could be shipped to them

for free (material condition) and asked to choose the op-

tion they would prefer. They were informed that one in

every 100 participants would be randomly selected to re-

ceive their chosen option. All options were valued at $50

(with the prize in the experiential condition being a $50

gift certificate for the chosen ice cream shop). The two

options within each pair were presented side-by-side,

with the order randomized. See appendix B for the

stimuli.

FIGURE 2

THE EFFECT OF EXPERIENTIAL (VS. MATERIAL) PURCHASE TYPE ON REVIEW RELIANCE IS MEDIATED BY QUALITY ASSESSMENT

BELIEFS (STUDY 3)

Experiential (vs. material) purchases

Quality assessments beliefs

Reliance on consumer reviews

-.42 (.16)*

.32 (.05)***

-.36 (.15)*

-.23 (.14)

NOTE.—Unstandardized regression coefﬁcients are shown, and standard errors are presented in parentheses. The coefﬁcient above the path from purchase type to re-

liance on consumer reviews represents the total effect without the mediator in the model; the coefﬁcient below the path represents the direct effect when the mediator

was included in the model. Coefﬁcients signiﬁcantly different from zero are indicated by asterisks (

p < .05,

p < .01,

***

p < .001).

5 When asked how many options were discussed in the five reviews

they read, seven participants did not write down a valid number; these

participants were excluded from the analysis for this question. Also,

one participant did not give a valid value for expected cost and was ex-

cluded from the analysis involving expected cost.

12 JOURNAL OF CONSUMER RESEARCH

Downloaded from https://academic.oup.com/jcr/advance-article-abstract/doi/10.1093/jcr/ucz042/5567101 by UCLA Digital Collections Services user on 20 February 2020

A pretest (N ¼ 54) that presented participants with the

same stimuli confirmed that the pair of ice cream shops

and the pair of ice cream makers were comparable in the

extent to which participants preferred one option over the

other. Specifically, 66.67% of participants in the experien-

tial condition and 70.37% of participants in the material

condition preferred the same option in the pair over the

other (v

(1) ¼ .09, p ¼ .77).

Reliance on Consumer Reviews. Participants were first

asked to choose one of the two presented options, with the

understanding that they had a chance of actually receiving

their selected option. After participants made their initial

selection, they saw a review ostensibly written by another

consumer about their selected option. All participants were

presented with the same slightly negative review: “When I

visited this ice cream shop [used this ice cream maker], I

was not very satisfied. I’m not sure I’d recommend it.”

Participants were then offered an opportunity to change

their mind and select the other option. The dependent mea-

sure was whether participants changed their mind and

switched to the other option after reading a negative review

about their initial choice.

Quality Assessment Beliefs. On the next page, partici-

pants were instructed to indicate the extent to which the re-

view reflected the consumer’s assessment of the objective

quality of the ice cream shop [ice cream maker] (1 ¼ “not

at all,” 9 ¼ “a great deal”), adapted from Spiller and

Belogolova (2017).

Manipulation Check. Participants rated the extent to

which a visit to an ice cream shop [an ice cream maker] is

material or experiential (1 ¼ “purely material,” 9 ¼

“purely experiential”).

Other Measures. Participants next rated how much

they cared about their decision, how important the choice

was to them, and how engaged they were in this decision

(1 ¼ “not at all,” 7 ¼ “very much”). We averaged these

three items to create a measure of purchase importance (a

¼ .88). Participants indicated how knowledgeable they

were about ice cream shops [ice cream makers] (1 ¼ “not

at all knowledgeable,” 7 ¼ “very knowledgeable”) and

whether they had previously heard of or visited [used] ei-

ther of the ice cream shops [ice cream makers]. In addition,

they rated the extent to which a visit to an ice cream shop

[an ice cream maker] was hedonic (1 ¼ “purely utili-

tarian,” 9 ¼ “purely hedonic”). See web appendix G for

the complete measures.

Results

Manipulation Check. Participants viewed a visit to an

ice cream shop as more experiential (M ¼ 6.21, SD ¼ 1.98)

than an ice cream maker (M ¼ 5.17, SD ¼ 2.01; t(216) ¼

3.84, p ¼ .0002, d ¼ .52).

Reliance on Consumer Reviews. Participants in the ex-

periential condition were less likely to change their mind

(34.95%) after reading the negative review than those in

the material condition (52.17%; v

(1) ¼ 6.54, p ¼ .01,

d ¼ .39

). This suggests that participants relied less on the

consumer review when deciding on an experiential pur-

chase than when deciding on a material purchase.

Mediation by Quality Assessment Beliefs. Participants

viewed the consumer review to be less reflective of the

consumer’s assessment of the option’s objective quality for

the experiential purchase (M ¼ 4.08, SD ¼ 2.11) than for

the material purchase (M ¼ 5.03, SD ¼ 1.91; t(216) ¼

3.52, p ¼ .0005, d ¼ .48). Moreover, a 5,000-sample boot-

strap analysis (model 4 in Hayes 2013) estimated an indi-

rect effect of purchase type on reliance on reviews via

quality assessment beliefs as –.30 (SE ¼ .12), and the 95%

bias-corrected CI of the indirect effect did not include zero

([–.58, –.12]). These results suggest that quality assessment

beliefs mediated the influence of purchase type on review

reliance (figure 3).

Other Measures. Purchase importance did not differ

between conditions (M

experiential

¼ 3.56, SD ¼ 1.31 vs.

material

¼ 3.43, SD ¼ 1.40; t(216) ¼ .74, p ¼ .46,

d ¼ .10). We did find that participants were more knowl-

edgeable about ice cream shops compared to ice cream

makers (M

experiential

¼ 2.87, SD ¼ 1.67 vs. M

material

1.83, SD ¼ 1.24; t(216) ¼ 5.30, p < .0001, d ¼ .72) and

rated a visit to an ice cream shop as more hedonic

experiential

¼ 7.54, SD ¼ 1.45 vs. M

material

¼ 6.65, SD ¼

1.77; t(216) ¼ 4.03, p ¼ .0001, d ¼ .55). Notably, however,

when we included these items in a multi-mediator model

(model 4 in Hayes 2013) along with quality assessment

beliefs, quality assessment beliefs was the only significant

mediator (indirect effect ¼ –.31, SE ¼ .05, 95% CI ¼

[–.61, –.11]).

Discussion

Employing a tightly controlled pair of experiential and

material purchases, a new behavioral measure of review re-

liance, and a virtually identical consumer review across

conditions, study 4 showed that people relied less on con-

sumer reviews when making an experiential purchase than

when making a material purchase. This effect was driven

by people’s view that consumer reviews of experiential

purchases are less likely to reflect objective quality than

consumer reviews of material purchases. Furthermore,

study 4 showed that alternative explanations (knowledge,

importance, and the hedonic nature of a purchase) were not

6 We conducted a logit regression to predict whether participants

chose the option with a positive review as a function of the experien-

tial (vs. material) manipulation. The binary indicator for the experien-

tial condition had an odds ratio of 0.49 (SE ¼ 0.14, p ¼ 0.01), which

can be translated into a Cohen’s d of –.39.

DAI, CHAN, AND MOGILNER 13

Downloaded from https://academic.oup.com/jcr/advance-article-abstract/doi/10.1093/jcr/ucz042/5567101 by UCLA Digital Collections Services user on 20 February 2020

responsible for the difference in review reliance between

purchase types, thus highlighting the critical role of quality

assessment beliefs. In the next study, we sought additional

evidence through a test of moderation that quality assess-

ment beliefs are responsible for the influence of purchase

type on review reliance.

STUDY 5: TEST OF MECHANISM

THROUGH MODERATION

The objective of study 5 was to provide further evidence

for the underlying role of quality assessment beliefs by ma-

nipulating whether reviews explicitly contained quality-

based assessments. If people believe that assessments of

experiential purchases are less based on objective quality

and this drives people to rely less on consumer reviews for

experiential purchases than material purchases, then the

difference in review reliance between purchase types

should decrease when a review for an experiential purchase

explicitly contains an assessment based on objective qual-

ity. This study followed a 2 (purchase type: experiential vs.

material)  2 (purchase assessment: control vs. quality)

between-subjects design.

Method

Participants. A total of 808 participants recruited

through MTurk completed this study in exchange for $.40.

As in studies 2 and 4, we excluded from the analysis 36

participants who had previously consumed the products. In

addition, 15 participants who reported that they could not

view the stimuli in the online study were excluded from

the analysis. The final sample included 757 participants

(45% female, 3 unspecified; M

age

¼ 36.1).

Purchase Type Manipulation. Participants were asked

to imagine that they were planning to go to an ice cream

shop (experiential condition) or buy an ice cream maker

(material condition), and that they were deciding between

two options of approximately the same price. Using the

same stimuli as in study 4 (see appendix B for stimuli), we

presented participants with either two options of ice cream

shops or two options of ice cream makers.

Quality Assessment Manipulation. Participants were

first asked to choose one of the two presented options.

After making their initial selection, participants saw a re-

view ostensibly written by another consumer about their

selected option. The structure of the review was the same

across all conditions, but the review in the quality condi-

tion explicitly assessed the objective quality of the option

(see appendix B for stimuli).

Reliance on Consumer Reviews. After being presented

with the slightly negative review, participants were asked

whether they would change their mind and choose the

other option (1 ¼ “definitely stick to my original choice,”

7 ¼ “definitely switch to the other ice cream shop [ice

cream maker]”). The likelihood of being influenced by the

review in their final decision served as the dependent mea-

sure of review reliance.

Manipulation Checks. On the next page, participants

were instructed to indicate the extent to which the review

they just read reflected the consumer’s assessment of the

objective quality of the ice cream shop [ice cream maker]

(1 ¼ “not at all,” 9 ¼ “a great deal”), adapted from Spiller

and Belogolova (2017). Participants also rated the extent to

which a visit to an ice cream shop [an ice cream maker] is

material or experiential (1 ¼ “purely material,” 9 ¼

“purely experiential”).

Other Measures. We used the same scales and meas-

ures as in study 4 to assess purchase importance (three

items; a ¼ .90), knowledge about ice cream shops [ice

cream makers], familiarity with the brands used in the

study, and the hedonic nature of the purchase. See web ap-

pendix H for the complete measures.

Results

Manipulation Checks. A 2 (purchase type)  2 (pur-

chase assessment) ANOVA on the quality assessment

FIGURE 3

THE EFFECT OF EXPERIENTIAL (VS. MATERIAL) PURCHASE TYPE ON REVIEW RELIANCE IS MEDIATED BY QUALITY ASSESSMENT

BELIEFS (STUDY 4)

Experiential (vs. material) purchase

Quality assessments beliefs

Reliance on a consumer review

-.96 (.27)***

.32 (.08)***

-.71 (.28)*

-.47 (.29)

NOTE.—Unstandardized regression coefﬁcients are shown, and standard errors are presented in parentheses. The coefﬁcient above the path from purchase type to re-

liance on consumer reviews represents the total effect without the mediator in the model; the coefﬁcient below the path represents the direct effect when the mediator

was included in the model. Coefﬁcients signiﬁcantly different from zero are indicated by asterisks (

p < .05,

p < .01,

***

p < .001). A logistic regression was used to

predict reliance on consumer reviews (a binary variable).

14 JOURNAL OF CONSUMER RESEARCH

Downloaded from https://academic.oup.com/jcr/advance-article-abstract/doi/10.1093/jcr/ucz042/5567101 by UCLA Digital Collections Services user on 20 February 2020

manipulation check revealed a main effect of the quality

assessment manipulation (M

quality

¼ 6.93, SD ¼ 1.67 vs.

control

¼ 5.50, SD ¼ 2.18; F(1, 753) ¼ 108.44, p <

.0001, g

¼ .13), a main effect of purchase type

experiential

¼ 5.74, SD ¼ 2.18 vs. M

material

¼ 6.67, SD ¼

1.85; F(1, 753) ¼ 44.35, p < .0001, g

¼ .06), and a sig-

nificant interaction (F(1, 753) ¼ 14.34, p ¼ .0002, g

.02). Consistent with our theory, participants in the control

condition reported that the review of the experiential pur-

chase was based less on quality (M ¼ 4.79, SD ¼ 2.13)

than the review of the material purchase (M ¼ 6.22, SD ¼

1.99; F(1, 753) ¼ 55.36, p < .0001, d ¼ .69), and this dif-

ference was attenuated in the quality condition (M

experiential

¼ 6.74, SD ¼ 1.75 vs. M

material

¼ 7.13, SD ¼ 1.58; F(1,

753) ¼ 4.07, p ¼ .04, d ¼ .24).

A two-way ANOVA on the experiential-material manip-

ulation check showed only the expected main effect of pur-

chase type (M

experiential

¼ 6.54, SD ¼ 2.09 vs. M

material

4.33, SD ¼ 2.21; F(1, 753) ¼ 200.95, p < .0001, g

.21). There was neither a significant main effect of the

quality manipulation (F(1, 753) ¼ 3.22, p ¼ .07, g

.004) nor an interaction (F(1, 753) ¼ .14, p ¼ .70, g

.0002).

Reliance on Consumer Reviews. A two-way ANOVA

on review reliance revealed a main effect of purchase type

experiential

¼ 4.28, SD ¼ 1.71 vs. M

material

¼ 4.72, SD ¼

1.47; F(1, 753) ¼ 13.72, p ¼ .0002, g

¼ .02), a main ef-

fect of quality assessment (M

quality

¼ 4.81, SD ¼ 1.56 vs.

control

¼ 4.20, SD ¼ 1.60; F(1, 753) ¼ 28.88, p < .0001,

¼ .04), and the predicted interaction (F(1, 753) ¼ 4.31,

p ¼ .038, g

¼ .01). Specifically, in the control condition,

participants relied less on the presented consumer review

when making an experiential purchase (M ¼ 3.87, SD ¼

1.66) than when making a material purchase (M ¼ 4.53, SD

¼ 1.47; F(1, 753) ¼ 16.95, p < .0001, d ¼ .42); however,

when the review explicitly assessed the option’s objective

quality, the difference between purchase types was not sta-

tistically significant (M

experiential

¼ 4.72, SD ¼ 1.65 vs.

material

¼ 4.90, SD ¼ 1.46; F(1, 753) ¼ 1.31, p ¼ .25,

d ¼ .12; figure 4).

Other Measures. A two-way ANOVA shows that pur-

chase type had a main effect on purchase importance

experiential

¼ 4.39, SD ¼ 1.45 vs. M

material

¼ 4.84, SD ¼

1.37; F(1, 753) ¼ 19.01, p < .0001, g

¼ .02), knowledge

experiential

¼ 3.86, SD ¼ 1.56 vs. M

material

¼ 2.75, SD ¼

1.64; F(1, 753) ¼ 91.84, p < .0001, g

¼ .11), and hedonic

ratings (M

experiential

¼ 7.76, SD ¼ 1.50 vs. M

material

¼ 6.34,

SD ¼ 2.19; F(1, 753) ¼ 107.36, p < .0001, g

¼ .12), and

the assessment manipulation influenced importance

quality

¼ 4.73, SD ¼ 1.46 vs. M

control

¼ 4.50, SD ¼ 1.39;

F(1, 753) ¼ 4.49, p ¼ .03, g

¼ .006) but not knowledge

or hedonic ratings (both ps > .27, g

< .002). Importantly,

there were no significant purchase type  assessment inter-

actions on any of these measures (all ps > .13, g

< .003).

Furthermore, when we included these measures as covari-

ates in a two-way ANCOVA model predicting review reli-

ance, the main effect of purchase type (F(1, 753) ¼ 5.40, p

¼ .02, g

¼ .01), the main effect of quality assessment

(F(1, 753) ¼ 25.44, p < .0001, g

¼ .03), and their interac-

tion (F(1, 753) ¼ 6.10, p ¼ .01, g

¼ .01) all remained sta-

tistically significant.

Discussion

As observed in our previous studies, study 5 showed that

participants were less willing to rely on a consumer review

when making an experiential purchase than when making a

material purchase. However, when the review explicitly

contained an assessment of the option’s objective quality,

people making an experiential purchase decision were as

likely to rely on the review as those deciding on a material

purchase. These findings provide further evidence for our

hypothesis that beliefs about the extent to which consumer

reviews contain assessments of objective quality are re-

sponsible for the lower reliance on consumer reviews ob-

served for experiential purchases (relative to material

purchases).

GENERAL DISCUSSION

Online consumer reviews have become a pervasive form

of social influence (Chen and Xie 2008). The present re-

search examines whether and why the type of purchase—

experiential or material—affects the extent to which people

rely on this prevalent source of information in their

decision-making.

Across one archival study, four experiments, and two ad-

ditional replications reported in the web appendix, this re-

search shows that although people often find reviews

useful for both experiential and material purchases, they

FIGURE 4

SHOPPERS’ RELIANCE ON CONSUMER REVIEWS AS A

FUNCTION OF PURCHASE TYPE AND ASSESSMENT OF

QUALITY (STUDY 5)

3.87

4.72

4.53

4.90

ytilauQlortnoC

Reliance on consumer reviews

eriential Material

DAI, CHAN, AND MOGILNER 15

Downloaded from https://academic.oup.com/jcr/advance-article-abstract/doi/10.1093/jcr/ucz042/5567101 by UCLA Digital Collections Services user on 20 February 2020

rely less on consumer reviews for experiential purchases

than for material purchases. This effect was robust across

millions of actual shoppers on Amazon (study 1), people

reading real consumer reviews they found online for a pur-

chase they intended to make (study 3), and people making

a consequential choice between two purchase options (stud-

ies 2 and 4). Across these studies, we observed the differ-

ence in review reliance for purchases that were relatively

more experiential or material at different points along the

material-experiential spectrum, including those closer to

the material end (study 1) or the middle or experiential end

(studies 2, 4, and 5). In particular, in study 3, the naturally

occurring purchases participants intended to make tended

to fell on one of the two ends of the material-experiential

spectrum, which allows us to replicate our effects with pro-

totypical material and experiential purchases.

This research also offers insight into the effect’s under-

lying mechanism. Specifically, people believe assessments

of experiential purchases are less based on objective qual-

ity than assessments of material purchases, and this belief

undermines people’s willingness to rely on other consum-

ers’ reviews for their experiential purchase decisions.

Studies 3 and 4 provided evidence for this mechanism

through mediation analyses, and study 5 provided evidence

through moderation. Furthermore, the studies ruled out al-

ternative explanations, such as purchase importance, pur-

chase knowledge, cost, and product desirability, through

stimuli selection, study design, and measurement.

Theoretical and Practical Implications

This research makes important contributions to the liter-

ature comparing experiential and material purchases.

Whereas the bulk of that literature has compared the down-

stream consequences of making experiential versus mate-

rial purchases (Gilovich et al. 2015a), the current findings

add to emerging work that tests pre-purchase differences.

While this emerging work has identified when and why

people might decide to make one type of purchase over the

other (Goodman et al. 2016; Goodman and Lim 2018;

Kumar et al. 2014; Kumar and Gilovich 2015, 2016;

Pchelin and Howell 2014; Tully et al. 2015), our research

further contributes by documenting a difference in how

people decide which option to buy within each purchase

type. Furthermore, by identifying the underlying role of

people’s beliefs about assessments as based on objective

quality, our research adds this variable to the list of per-

ceived differences—such as comparability, interchange-

ability, personal relevance, feelings of gratitude, and

emotional intensity (Carter and Gilovich 2010, 2012; Chan

and Mogilner 2017; Rosenzweig and Gilovich 2012;

Walker et al. 2016)—that psychologically distinguishes ex-

periential and material purchases.

This research also contributes to the body of research ex-

amining people’s willingness to rely on consumer reviews

in two ways (Chen and Lurie 2013; de Langhe et al. 2016;

Moore 2015; Naylor, Lamberton, and Norton 2011; Yin

et al. 2014, 2017). First, our findings reveal the critical role

of quality assessment beliefs in affecting review reliance.

Second, these findings identify the experiential versus ma-

terial nature of a purchase as a distinct and novel delinea-

tion across product domains that determines reliance on

consumer reviews. This categorization is distinct from a

product’s hedonic versus utilitarian nature, which has also

been shown to influence review helpfulness (Chu et al.

2015; Moore 2015; Sen and Lerman 2007). We show that

the experiential-material effect persisted when we

controlled for the purchase’s hedonic (vs. utilitarian) nature

(studies 1, 3, 4, and 5), and when the experiential purchase

was viewed as less hedonic than the material purchase

(study 2). This categorization is also distinct from the de-

lineation between search goods and experience goods,

which too has been shown to influence the impact of con-

sumer reviews (Mudambi and Schuff 2010; Park and Lee

2009). While the experiential (vs. material) distinction cen-

ters on consumers’ primary purpose of making a purchase

(i.e., gaining an experience vs. acquiring a possession; Van

Boven and Gilovich 2003), the search (vs. experience)

goods distinction reflects how easily consumers can evalu-

ate a product prior to consumption (Huang, Lurie, and

Mitra 2009; Nelson 1970). Not only is the difference be-

tween search and experience goods blurred in the online

context our research examines (Huang et al. 2009), but we

replicated our effect controlling for differences in evalu-

ability by having participants focus on the experiential (vs.

material) aspects of the very same purchase (see the repli-

cation experiments in web appendixes D and E). The cur-

rent research thus suggests a new way of slicing the

consumer product landscape to identify the types of prod-

ucts that are more likely to benefit from word of mouth.

In addition, this research contributes to more recent re-

search on quality assessment beliefs (Spiller and

Belogolova 2017) by showing that beliefs about quality

assessments systematically differ across product domains

and affect people’s susceptibility to social influence. Our

results may also provide new insight into the construct. In

Spiller and Belogolova (2017), participants were given the

opportunity to justify their choice of option in one of three

ways by reporting: a) the chosen option was objectively

better than the forgone option, b) their choice was a matter

of taste, or c) they did not have enough knowledge to

judge. This implies that if people believe a purchase is pri-

marily assessed based on objective quality, it is not based

on taste. Interestingly, however, when we asked partici-

pants in the pilot study and study 3 to separately rate the

extent to which they believed assessments of a purchase

are (1) a matter of objective quality and (2) a matter of

taste, we did not observe a significant negative correlation

between these items. Moreover, though participants be-

lieved that assessments of experiential purchases were

16 JOURNAL OF CONSUMER RESEARCH

Downloaded from https://academic.oup.com/jcr/advance-article-abstract/doi/10.1093/jcr/ucz042/5567101 by UCLA Digital Collections Services user on 20 February 2020

more based on taste than material purchases, such beliefs

did not significantly predict review reliance (see web ap-

pendixes A and F). We speculate this is because people

tend to assume that other consumers share their tastes

(Naylor et al. 2011) and thus may not discount reviews

even if reviews contain taste-based assessments. Future re-

search should more systematically examine the relationship

between taste and quality assessment beliefs and compare

their roles in social influence.

The current research offers a number of practical impli-

cations for firms and review websites. For instance, our

results suggest that firms should take their product type

(experiential vs. material) into account when designing

their optimal communication strategy. In an additional ex-

periment (N ¼ 243 MTurk participants, 40% female, M

age

¼ 30.6), we found that whereas people relied less on con-

sumer reviews for an experiential purchase (a Broadway

show ticket) than a material purchase (a pair of speakers),

there was no significant difference in how useful people

considered company-provided information (web appendix

I). Noting that people do not discount all information more

for experiential purchases, firms could leverage these

insights to emphasize whichever would be the more per-

suasive source of information for their products. Further,

retailers may be able to dampen the impact of a negative

consumer review by highlighting the experiential aspects

of their product. In addition, this research advises review

websites (particularly those featuring experiential pur-

chases) to address users’ doubts about another consumer’s

evaluation as bearing on their own by helping users iden-

tify reviews that provide quality-based judgments or by

encouraging review writers to explicitly include quality-

based assessments in their reviews.

Future Directions

These findings may lead readers to wonder about the

popularity of websites that feature consumer reviews for

experiential purchases. Indeed, websites such as Yelp and

TripAdvisor have no shortage of engaged and active users.

Importantly, the current findings do not suggest that people

do not visit review websites or read reviews for experien-

tial purchases. Rather, this research examined the extent to

which people find the reviews they read useful and are

influenced by them. More precisely, this research compares

the extent to which people rely on consumer reviews for

experiential purchases relative to material purchases during

their decision process. Indeed, in our studies, we observe

that people generally consider consumer reviews useful for

both experiential and material purchases: the average re-

view reliance was significantly above the scale midpoint

for both experiential and material purchases in studies 1, 2,

and 3 (ps < .002). However, we also observe across all of

the studies that people rely less on consumer reviews for

experiences than for material goods.

Moreover, there are other factors—such as consumers’

motivations to write reviews—that might also contribute to

the popularity of review websites featuring experiential

purchases. Self-enhancement and the desire to converse are

two important drivers of individuals’ motivation to spread

word of mouth (Lovett, Peres, and Shachar 2013). Since

people are judged more positively when talking about their

experiential purchases than their material purchases (Van

Boven, Campbell, and Gilovich 2010) and derive greater

enjoyment from doing so (Kumar and Gilovich 2015), peo-

ple may be more motivated to write and post a review

about their recently acquired experience than about their

newly acquired possession. Also, consumers are more

likely to express gratefulness in their reviews for experien-

ces than for material possessions (Walker et al. 2016).

Future research should systematically explore the extent to

which activity on review websites is driven by the motiva-

tions of the consumers writing the reviews.

At first glance, the findings documented in this research

seem to counter those showing that conversation partners

enjoy hearing more about each other’s recent experiential

purchase than each other’s recent material purchase (Van

Boven et al. 2010). Importantly, however, people likely

have different motives when reading a review from an

anonymous consumer to inform a purchase decision than

when meeting someone in person and hearing his or her

story about a recent purchase. Whereas people read con-

sumer reviews with the primary intention to use those

reviews to predict their own evaluations of a product, peo-

ple engage in conversation with the primary intention to

connect with others. Thus, the reasons people more easily

connect with others who tell stories about their experiential

purchases (Van Boven et al. 2010)—experiential purchases

evoke stronger emotions (Chan and Mogilner 2017), are

less stigmatized as materialistic (Van Boven et al. 2010),

and are more revealing of the storyteller’s sense of self

(Carter and Gilovich 2010)—are not factors that would ob-

viously contribute to the usefulness of consumer reviews.

Still, future research could examine how shoppers’ social

motives influence their reliance on others’ opinions, in-

cluding those of anonymous online reviewers, in-store

sales associates, and friends.

Study 4 showed that even when the review content was

exactly the same, people perceived the review to be less re-

flective of the option’s objective quality when the purchase

was experiential (vs. material). This suggests that it may

not just be the content of the reviews, but rather shoppers’

beliefs about the basis of other consumers’ assessments that

underlie the observed effect on review reliance. This poses

the interesting question of whether it is appropriate for peo-

ple to discount the value of consumer reviews for experien-

tial purchases (relative to material purchases). Notably,

people’s lay beliefs about the alignment between their own

preferences and those of a stranger are not always accurate

(Barasz, Kim, and John 2016; Naylor et al. 2011), and

DAI, CHAN, AND MOGILNER 17

Downloaded from https://academic.oup.com/jcr/advance-article-abstract/doi/10.1093/jcr/ucz042/5567101 by UCLA Digital Collections Services user on 20 February 2020

people underestimate the value of knowing others’ reactions

to an event in predicting their own reactions to that event

(Eggleston et al. 2015; Gilbert et al. 2009; Mu¨ller-Trede

et al. 2018). So, do people underestimate the value of con-

sumer reviews for experiential purchases? An exploratory

study that we conducted suggests so (web appendix J). In

the study, half of the participants simply read a product re-

view written by another participant who had tried the prod-

uct (a bag of chips in the experiential condition or a

flashlight in the material condition). These participants con-

sidered the review of the experiential purchase to be less

useful than the review of the material purchase (as found in

studies 1–5). The other half of participants actually tried the

product prior to evaluating it and reading a review about it;

the discrepancy in product evaluations between these partic-

ipants and their corresponding review writers did not signifi-

cantly differ between the experiential and material

conditions. This observation provides preliminary evidence

to suggest that people underestimate how much consumer

reviews can predict their enjoyment of experiential pur-

chases (relative to material purchases). It would be interest-

ing for future research to identify whether people’s

reluctance to rely on reviews for experiential purchases (rel-

ative to material goods) might lead to inferior purchase deci-

sions (Fitzsimons and Lehmann 2004).

Study 5 showed that people rely on reviews for experi-

ential and material purchases equally when the reviews ex-

plicitly involve quality judgments. That is, when it is

salient to shoppers that consumers’ judgments of experien-

tial purchases reflect objective quality, shoppers are more

willing to rely on consumer reviews for experiential pur-

chases. Future research could explore other cases in which

such perceptions are heightened, perhaps from aggregated

consumer ratings (e.g., average star ratings; de Langhe

et al. 2016) or when the text of reviews contains minimal

self-referencing language (Spiller and Belogolova 2017).

Gaining a broader understanding of when people rely more

on reviews for experiential purchases would be valuable.

Finally, future research might explore how characteris-

tics of reviews and product type jointly influence review

helpfulness. For instance, we conducted a post hoc analysis

to examine the role of review valence using Amazon

reviews (study 1). The positive and significant interaction

between star rating and experiential rating (p < .001; web

appendix B) suggests that the differences in review help-

fulness between experiential and material purchases may

be smaller for positive reviews than for negative reviews.

We note, however, that the negative relationship between

experiential ratings and review helpfulness remained statis-

tically significant at each star rating level (all ps < .001),

which suggests that the negative effect of experiential (vs.

material) purchase type on review helpfulness holds for

both positive and negative reviews. Research could simi-

larly examine how experiential-material purchase type

interacts with other well-established purchase categoriza-

tions. For example, we explored whether the hedonic (vs.

utilitarian) nature of a purchase moderates the effect of ex-

periential (vs. material) purchase type on review reliance,

but we did not find consistent results for the interaction

across studies 1, 4, and 5 (see web appendixes B, G, and

H). Future research that systematically investigates the in-

teraction between review characteristics and various types

of purchases would not only be theoretically interesting,

but also practically meaningful to marketers who worry

about the impact of negative reviews while wanting to le-

verage positive reviews.

Conclusions

This research highlights that experiential and material

purchases differ not only in their likelihood of being con-

sumed or in their enjoyment once consumed, but also in

the decision processes through which people choose which

option to buy. Our findings reveal that people believe

assessments of experiential purchases to be less a matter of

objective quality than assessments of material purchases.

Thus, people rely less on consumer reviews when deciding

among experiences than when deciding among material

goods, thereby suggesting that people are less receptive to

advice on what to do than on what to have.

DATA COLLECTION INFORMATION

Study 1 relied on data that was collected and made

publicly available by other researchers (McAuley and

Leskovec 2013; available at https://snap.stanford.edu/

data/web-Amazon.html). The first author cleaned the

original data to create a dataset appropriate for this

article, and all authors jointly interpreted the data. For the

experimental studies (studies 2–5), all authors supervised

data collection, and the first author conducted analyses.

Study 2 was run by research assistants at the University

of Toronto in summer 2015 among participants recruited

on campus and through the university behavioral lab.

Studies 3 and 5 were conducted online using Amazon

Mechanical Turk panelists and were run in spring 2018.

Study 4 was conducted online using the subject pool of

the Behavioral Lab at Anderson School of Management

in spring 2018. Data from laboratory experiments are

available at https://osf.io/b25tm/.

18 JOURNAL OF CONSUMER RESEARCH

Downloaded from https://academic.oup.com/jcr/advance-article-abstract/doi/10.1093/jcr/ucz042/5567101 by UCLA Digital Collections Services user on 20 February 2020

APPENDIX A

Study Stimuli (Study 2)

Below are the pictures, product descriptions, and consumer reviews used in study 2. We randomized which side (left or

right) each option appeared, as well as which of the options received a positive review. One display order is shown here as

an example.

EXPERIENTIAL CONDITION

DAI, CHAN, AND MOGILNER 19

Downloaded from https://academic.oup.com/jcr/advance-article-abstract/doi/10.1093/jcr/ucz042/5567101 by UCLA Digital Collections Services user on 20 February 2020

MATERIAL CONDITION

20 JOURNAL OF CONSUMER RESEARCH

Downloaded from https://academic.oup.com/jcr/advance-article-abstract/doi/10.1093/jcr/ucz042/5567101 by UCLA Digital Collections Services user on 20 February 2020

APPENDIX B

Study Stimuli (Studies 4 and 5)

Below are the pictures and consumer reviews

used in studies 4 and 5. We randomized which side (left or right) each

option appeared. One display order is shown here as an example.

7 When developing the stimuli, our goal was to make the review in the control condition less explicitly based on judgments of quality than the review

in the quality condition. Since individuals differ in their preference for variety (Steenkamp and Baumgartner 1992; Van Trijp et al. 1996), which sug-

gests that a greater amount of variety is not objectively better than less variety and that variety is not considered a dimension of quality, we included

mention of this attribute in the review for the control condition. The manipulation check confirmed that our manipulation was effective.

DAI, CHAN, AND MOGILNER 21

Downloaded from https://academic.oup.com/jcr/advance-article-abstract/doi/10.1093/jcr/ucz042/5567101 by UCLA Digital Collections Services user on 20 February 2020

REFERENCES

Ante, Spencer E. (2009), “Amazon: Turning Consumer Opinions

into Gold,” Bloomberg Businessweek, October 15, https://

www.bloomberg.com/news/articles/2009-10-15/amazon-

turning-consumer-opinions-into-gold.

Barasz, Kate, Tami Kim, and Leslie K. John (2016), “The Role of

(Dis)Similarity in (Mis)Predicting Others’ Preferences,”

Journal of Marketing Research, 53 (4), 597–607.

Berger, Jonah (2014), “Word of Mouth and Interpersonal

Communication: A Review and Directions for Future

Research,” Journal of Consumer Psychology, 24 (4),

586–607.

Caprariello, Peter A. and Harry T. Reis (2013), “To Do, to Have,

or to Share? Valuing Experiences over Material Possessions

Depends on the Involvement of Others,” Journal of

Personality and Social Psychology, 104 (2), 199–215.

Carter, Travis J. and Thomas Gilovich (2010), “The Relative

Relativity of Material and Experiential Purchases,” Journal

of Personality and Social Psychology, 98 (1), 146–59.

——— (2012), “I Am What I Do, Not What I Have: The

Differential Centrality of Experiential and Material

Purchases to the Self,” Journal of Personality and Social

Psychology, 102 (6), 1304–17.

Chan, Cindy and Cassie Mogilner (2017), “Experiential Gifts

Foster Stronger Social Relationships than Material Gifts,”

Journal of Consumer Research, 43 (6), 913–31.

Chen, Pei-Yu, Samita Dhanasobhon, and Michael D. Smith

(2008), “All Reviews Are Not Created Equal: The

Disaggregate Impact of Reviews and Reviewers at

Amazon.com,” https://dx.doi.org/10.2139/ssrn.918083.

Chen, Yubo and Jinhong Xie (2008), “Online Consumer Review:

Word-of-Mouth as a New Element of Marketing

Communication Mix,” Management Science, 54 (3), 477–91.

Chen, Zoey and Nicholas H. Lurie (2013), “Temporal Contiguity

and Negativity Bias in the Impact of Online Word of Mouth,”

Journal of Marketing Research, 50 (4), 463–76.

Chevalier, Judith A. and Dina Mayzlin (2006), “The Effect of

Word of Mouth on Sales: Online Book Reviews,” Journal of

Marketing Research, 43 (3), 345–54.

Chu, Wujin, Minjung Roh, and Kiwan Park (2015), “The Effect of

the Dispersion of Review Ratings on Evaluations of Hedonic

versus Utilitarian Products,” International Journal of

Electronic Commerce, 19 (2), 95–125.

Cialdini, Robert B. and Noah J. Goldstein (2004), “Social

Inﬂuence: Compliance and Conformity,” Annual Review of

Psychology, 55, 591–621.

de Langhe, Bart, Philip M. Fernbach, and Donald R. Lichtenstein

(2016), “Navigating by the Stars: Investigating the Actual

and Perceived Validity of Online User Ratings,” Journal of

Consumer Research, 42 (6), 817–33.

Eggleston, Casey M., Timothy D. Wilson, Minha Lee, and Daniel

T. Gilbert (2015), “Predicting What We Will Like: Asking a

Stranger Can Be as Good as Asking a Friend,”

Organizational Behavior and Human Decision Processes,

128, 1–10.

Eliashberg, Jehoshua and Mohanbir S. Sawhney (1994),

“Modeling Goes to Hollywood: Predicting Individual

Differences in Movie Enjoyment,” Management Science,40

(9), 1151–73.

Feick, Lawrence and Robin A. Higie (1992), “The Effects of

Preference Heterogeneity and Source Characteristics on Ad

Processing and Judgments about Endorsers,” Journal of

Advertising, 21 (2), 9–24.

Fitzsimons, Gavan J. and Donald R. Lehmann (2004), “Reactance

to Recommendations: When Unsolicited Advice Yields

Contrary Responses,” Marketing Science, 23 (1), 82–94.

Forman, Chris, Anindya Ghose, and Batia Wiesenfeld (2008),

“Examining the Relationship between Reviews and Sales:

The Role of Reviewer Identity Disclosure in Electronic

Markets,” Information Systems Research, 19 (3), 291–313.

Gilbert, Daniel T., Matthew A. Killingsworth, Rebecca N. Eyre,

and Timothy D. Wilson (2009), “The Surprising Power of

Neighborly Advice,” Science, 323 (5921), 1617–9.

Gilovich, Thomas, Amit Kumar, and Lily Jampol (2015a), “A

Wonderful Life: Experiential Consumption and the Pursuit of

Happiness,” Journal of Consumer Psychology, 25 (1), 152–65.

———. (2015b), “The Beach, the Bikini, and the Best Buy: Replies

to Dunn and Weidman, and to Schmitt, Brakus, and

Zarantonello,” Journal of Consumer Psychology, 25 (1),

179–84.

Godes, David and Dina Mayzlin (2004), “Using Online

Conversations to Study Word-of-Mouth Communication,”

Marketing Science, 23 (4), 545–60.

Goldstein, Noah J., Robert B. Cialdini, and Vladas Griskevicius

(2008), “A Room with a Viewpoint: Using Social Norms to

Motivate Environmental Conservation in Hotels,” Journal of

Consumer Research, 35 (3), 472–82.

Goodman, Joseph K. and Sarah Lim (2018), “When Consumers

Prefer to Give Material Gifts Instead of Experiences: The

Role of Social Distance,” Journal of Consumer Research,45

(2), 365–82.

Goodman, Joseph K., Selin A. Malkoc, and Brittney L.

Stephenson (2016), “Celebrate or Commemorate? A Material

Purchase Advantage When Honoring Special Life Events,”

Journal of the Association for Consumer Research, 1 (4),

497–508.

Gorenﬂo, Daniel W. and William D. Crano (1989), “Judgmental

Subjectivity/Objectivity and Locus of Choice in Social

Comparison,” Journal of Personality and Social Psychology,

57 (4), 605–14.

Hayes, Andrew F. (2013), Introduction to Mediation, Moderation,

and Conditional Process Analysis: A Regression-Based

Approach, New York: Guilford.

Holbrook, Morris B. and Elizabeth C. Hirschman (1982), “The

Experiential Aspects of Consumption: Consumer Fantasies,

Feelings, and Fun,” Journal of Consumer Research, 9 (2),

132–40.

Hong, Pat (2015), “The Amazon Customer Journey: Leading the

Way for Consumer-Centric Ecommerce,” Linkdex, May 21,

https://www.linkdex.com/en-us/inked/amazon-customer-

journey/.

Huang, Peng, Nicholas H. Lurie, and Sabyasachi Mitra (2009),

“Searching for Experience on the Web: An Empirical

Examination of Consumer Behavior for Search and

Experience Goods,” Journal of Marketing, 73 (2), 55–69.

Johansson, Johny K., Susan P. Douglas, and Ikujiro Nonaka

(1985), “Assessing the Impact of Country of Origin on

Product Evaluations: A New Methodological Perspective,”

Journal of Marketing Research, 22 (4), 388–96.

Keen, Andrew (2008), The Cult of the Amateur: How Blogs,

MySpace, YouTube, and the Rest of Today’s User-Generated

Media Are Destroying Our Economy, Our Culture, and Our

Values, New York: Crown.

22 JOURNAL OF CONSUMER RESEARCH

Downloaded from https://academic.oup.com/jcr/advance-article-abstract/doi/10.1093/jcr/ucz042/5567101 by UCLA Digital Collections Services user on 20 February 2020

Kumar, Amit and Thomas Gilovich (2015), “Some ‘Thing’ to

Talk About? Differential Story Utility from Experiential and

Material Purchases,” Personality and Social Psychology

Bulletin, 41 (10), 1320–31.

———. (2016), “To Do or to Have, Now or Later? The Preferred

Consumption Proﬁles of Material and Experiential

Purchases,” Journal of Consumer Psychology, 26 (2),

169–78.

Kumar, Amit, Matthew A. Killingsworth, and Thomas Gilovich

(2014), “Waiting for Merlot: Anticipatory Consumption of

Experiential and Material Purchases,” Psychological

Science, 25 (10), 1924–31.

Lovett, Mitchell J., Renana Peres, and Ron Shachar (2013), “On

Brands and Word of Mouth,” Journal of Marketing

Research, 50 (4), 427–44.

Marchant, Ross (2015), “Local Consumer Review Survey 2015,”

BrightLocal, October 19, https://www.brightlocal.com/re-

search/local-consumer-review-survey-2015/.

Mayzlin, Dina, Yaniv Dover, and Judith A. Chevalier (2012),

“Promotional Reviews: An Empirical Investigation of Online

Review Manipulation,” NBER Working Paper No. 18340,

National Bureau of Economic Research, Cambridge, MA

02138.

McAuley, Julian and Jure Leskovec (2013), “Hidden Factors and

Hidden Topics: Understanding Rating Dimensions with

Review Text,” in Proceedings of the 7th ACM Conference on

Recommender Systems, ed. Qiang Yang, Irwin King, and

Qing Li, New York: ACM, 165–72.

Moe, Wendy W. and Michael Trusov (2011), “The Value of

Social Dynamics in Online Product Ratings Forums,”

Journal of Marketing Research, 48 (3), 444–56.

Moore, Sarah G. (2015), “Attitude Predictability and Helpfulness

in Online Reviews: The Role of Explained Actions and

Reactions,” Journal of Consumer Research, 42 (1), 30–44.

Mudambi, Susan M. and David Schuff (2010), “What Makes a

Helpful Online Review? A Study of Customer Reviews on

Amazon.com,” MIS Quarterly, 34 (1), 185–200.

Mu¨ller-Trede, Johannes, Shoham Choshen-Hillel, Meir Barneron,

and Ilan Yaniv (2018), “The Wisdom of Crowds in Matters

of Taste,” Management Science, 64 (4), 1779–973.

Naylor, Rebecca W., Cait P. Lamberton, and David A. Norton

(2011), “Seeing Ourselves in Others: Reviewer Ambiguity,

Egocentric Anchoring, and Persuasion,” Journal of

Marketing Research, 48 (3), 617–31.

Nelson, Phillip (1970), “Information and Consumer Behavior,”

Journal of Political Economy, 78 (2), 311–29.

Nicolao, Leonardo, Julie R. Irwin, and Joseph K. Goodman

(2009), “Happiness for Sale: Do Experiential Purchases

Make Consumers Happier than Material Purchases?” Journal

of Consumer Research, 36 (2), 188–98.

Nielsen (2015), “Recommendations from Friends Remain Most

Credible Form of Advertising among Consumers; Branded

Websites Are the Second-Highest-Rated Form,” September 28,

http://www.nielsen.com/ca/en/press-room/2015/recommendations-

from-friends-remain-most-credible-form-of-advertising.html.

Olson, James M., Robert J. Ellis, and Mark P. Zanna (1983),

“Validating Objective versus Subjective Judgments:

Interest in Social Comparison and Consistency

Information,” Personality and Social Psychology Bulletin,

9 (3), 427–36.

Park, Cheol and Thae Min Lee (2009), “Information Direction,

Website Reputation and eWOM Effect: A Moderating Role of

Product Type,” Journal of Business Research, 62 (1), 61–7.

Pchelin, Paulina and Ryan T. Howell (2014), “The Hidden Cost of

Value-Seeking: People Do Not Accurately Forecast the

Economic Beneﬁts of Experiential Purchases,” Journal of

Positive Psychology, 9 (4), 322–34.

Price, Linda L., Lawrence F. Feick, and Robin A. Higie (1989),

“Preference Heterogeneity and Coorientation as

Determinants of Perceived Informational Inﬂuence,” Journal

of Business Research, 19 (3), 227–42.

Rosenzweig, Emily and Thomas Gilovich (2012), “Buyer’s

Remorse or Missed Opportunity? Differential Regrets for

Material and Experiential Purchases,” Journal of Personality

and Social Psychology, 102 (2), 215–23.

Sen, Shahana and Dawn Lerman (2007), “Why Are You Telling

Me This? An Examination into Negative Consumer Reviews

on the Web,” Journal of Interactive Marketing, 21 (4),

76–94.

Simonson, Itamar and Emanuel Rosen (2014), Absolute Value:

What Really Inﬂuences Customers in the Age of (Nearly)

Perfect Information, New York: HarperCollins.

Spears, Russell, Naomi Ellemers, and Bertjan Doosje (2009),

“Strength in Numbers or Less Is More? A Matter of Opinion

and a Question of Taste,” Personality and Social Psychology

Bulletin, 35 (8), 1099–111.

Spiller, Stephen A. and Lena Belogolova (2017), “On Consumer

Beliefs about Quality and Taste,” Journal of Consumer

Research, 43 (6), 970–91.

Steenkamp, Jan-Benedict E. M. and Hans Baumgartner (1992),

“The Role of Optimum Stimulation Level in Exploratory

Consumer Behavior,” Journal of Consumer Research, 19 (3),

434–48.

Tirole, Jean (1988), The Theory of Industrial Organization,

Cambridge, MA: MIT Press.

Tirunillai, Seshadri and Gerard J. Tellis (2012), “Does Chatter

Really Matter? Dynamics of User-Generated Content and

Stock Performance,” Marketing Science, 31 (2), 198–215.

Tully, Stephanie M., Hal E. Hershﬁeld, and Tom Meyvis (2015),

“Seeking Lasting Employment with Limited Money:

Financial Constraints Increase Preference for Material Goods

over Experiences,” Journal of Consumer Research, 42 (1),

59–75.

Van Boven, Leaf, Margaret C. Campbell, and Thomas Gilovich

(2010), “Stigmatizing Materialism: On Stereotypes and

Impressions of Materialistic and Experiential Pursuits,”

Personality and Social Psychology Bulletin,36(4),

551–63.

Van Boven, Leaf and Thomas Gilovich (2003), “To Do or to

Have? That Is the Question,” Journal of Personality and

Social Psychology, 85 (6), 1193–202.

Van Trijp, Hans C. M., Wayne D. Hoyer, and J. Jeffrey Inman

(1996), “Why Switch? Product Category: Level Explanations

for True Variety-Seeking Behavior,” Journal of Marketing

Research, 33 (3), 281–92.

Walker, Jesse, Amit Kumar, and Thomas Gilovich (2016),

“Cultivating Gratitude and Giving through Experiential

Consumption,” Emotion, 16 (8), 1126–36.

Yaniv, Ilan, Shoham Choshen-Hillel, and Maxim Milyavsky

(2011), “Receiving Advice on Matters of Taste: Similarity,

DAI, CHAN, AND MOGILNER 23

Downloaded from https://academic.oup.com/jcr/advance-article-abstract/doi/10.1093/jcr/ucz042/5567101 by UCLA Digital Collections Services user on 20 February 2020

Majority Inﬂuence, and Taste Discrimination,” Organizational

Behavior and Human Decision Processes, 115 (1), 111–20.

Yin, Dezhi, Samuel Bond, and Han Zhang (2014), “Anxious or

Angry? Effects of Discrete Emotions on the Perceived

Helpfulness of Online Reviews,” MIS Quarterly, 38 (2),

539–60.

———. (2017), “Keep Your Cool or Let It Out: Nonlinear Effects

of Expressed Arousal on Perceptions of Consumer Reviews,”

Journal of Marketing Research, 54 (3), 447–63.

Zeithaml, Valarie A. (1988), “Consumer Perceptions of Price,

Quality, and Value: A Means-End Model and Synthesis of

Evidence,” Journal of Marketing, 52 (3), 2–22.

24 JOURNAL OF CONSUMER RESEARCH

Downloaded from https://academic.oup.com/jcr/advance-article-abstract/doi/10.1093/jcr/ucz042/5567101 by UCLA Digital Collections Services user on 20 February 2020